🔬 mesoscale physics

Supercurrent tuning of the Josephson coupling energy

This paper proposes a multiterminal device that tunes Josephson coupling energy via supercurrent bias from separate leads, offering a flux-loop-free alternative for frequency-tunable qubits with significantly reduced susceptibility to magnetic noise.

Original authors: Maxwell Wisne, Venkat Chandrasekhar

Published 2026-02-26

📖 4 min read☕ Coffee break read

CC BY 4.0

Original authors: Maxwell Wisne, Venkat Chandrasekhar

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

The Big Idea: Tuning a Radio Without the Antenna

Imagine you have a very sensitive radio (a superconducting qubit) that needs to be tuned to a specific frequency to work. In the world of quantum computers, this "tuning" is done by adjusting the Josephson Energy, which is essentially the "stiffness" or strength of the connection between two superconducting wires.

The Old Way (The Problem):
Traditionally, to tune this radio, scientists used a magnetic loop (like a SQUID). Think of this like trying to tune a radio by waving a giant, noisy magnet around it.

The Issue: While this works, that big magnet acts like a radio antenna. It picks up every little bit of magnetic "static" or noise from the environment (like Wi-Fi signals, power lines, or even the Earth's magnetic field). This noise makes the quantum computer unstable and causes errors.

The New Way (The Solution):
In this paper, the researchers at Northwestern University found a clever way to tune the radio without the big magnet. Instead of using a magnetic loop, they used a second superconducting wire to "push" on the first one.

They call this a multiterminal device. Imagine a four-way intersection:

The Sample Junction: The main road where the traffic (supercurrent) flows. This is the part we want to tune.
The Control Junction: A side road that feeds into the main road.

By sending a specific amount of "traffic" (supercurrent) down the side road, they can change how the main road behaves. It's like having a traffic controller on a side street who can slow down or speed up the cars on the main highway just by changing the timing of their own lights, without ever needing to bring in a giant, noisy construction crew (the magnetic field).

The Key Findings

1. The "Volume Knob" Effect
The researchers discovered that as they increased the supercurrent on the "side road" (the control junction), the "stiffness" (critical current) of the main road decreased smoothly.

Analogy: Imagine you are pushing a heavy swing. If someone else starts pushing against you from the side, it becomes harder for you to push the swing high. The more they push back, the lower the swing goes. They found they could lower the "swing height" (the energy) of the quantum bit by about 20% just by adjusting this side current.

2. No More Magnetic Noise
Because they aren't using a magnetic loop to do the tuning, the quantum bit is no longer exposed to that magnetic "static."

Analogy: It's like switching from a radio that picks up static from a nearby power line to a radio that is hard-wired directly to the station. The signal is much cleaner. This means the quantum computer could be more stable and make fewer mistakes.

3. The Shape of the Wave
They also noticed that the relationship between the current and the energy wasn't a perfect, smooth wave (like a sine wave). It was a bit wobbly and complex.

Analogy: If a normal wave is like a perfect ocean swell, their wave was like a choppy sea with extra bumps. This "wobble" actually gives them more control, allowing them to shape the quantum bit's behavior in unique ways that a simple magnetic loop couldn't do.

Why Does This Matter?

Quantum computers are incredibly fragile. They are like glass houses in a storm; even a tiny bit of noise can break them.

Current State: We have to use magnetic loops to tune them, but those loops let the "storm" (noise) in.
Future State: This new method allows us to tune the computer using electricity (current) instead of magnetism. It's like building a soundproof room for the quantum bit.

In a nutshell: The researchers built a new type of switch that lets them tune a quantum computer's frequency using a "push" from a neighboring wire, rather than a "pull" from a noisy magnet. This could lead to much more reliable and powerful quantum computers in the future.

1. Problem Statement

Superconducting qubits rely on Josephson junctions (JJs) as nonlinear inductors, where the qubit frequency is determined by the Josephson coupling energy ( $E_J$ ).

Current Limitation: In frequency-tunable qubits, $E_J$ is typically tuned by threading magnetic flux through a Superconducting Quantum Interference Device (SQUID) loop containing two junctions. This approach exposes the qubit to magnetic environmental noise (flux noise), which degrades coherence.
Alternative Limitations: While "gatemon" qubits (using gate-tunable SNS junctions) offer electrical tuning, they often require specific materials (like graphene) or oxide layers that can introduce loss.
Goal: The authors aim to demonstrate a method to tune the critical current ( $I_c$ ) and thus $E_J$ of a single Josephson junction without using a magnetic flux loop, thereby reducing susceptibility to flux noise.

2. Methodology

The researchers fabricated and characterized multiterminal Josephson junctions (4TJJ) and SQUID loops using a diffusive normal metal (Gold, Au) weak link coupled to superconducting leads (Aluminum, Al).

Device Geometry:
- 4TJJ: A four-terminal device with a "sample" junction (leads 1 & 2) and a "control" junction (leads 3 & 4). The control leads attach to the sample junction at 45-degree angles.
- SQUID: Two SNS junctions connected in parallel to form a loop, with additional leads on one side to apply bias.
Fabrication: Devices were made on Si substrates with SiO $_2$ $_{2}$ using e-beam lithography. A specific process was used to maximize superconducting coupling:
- Sequential evaporation of Au and Al.
- In-situ plasma cleaning (O $_2$ and Ar $^+$ ) to ensure high-quality interfaces.
- Rapid cooling (<3 hours post-deposition) to prevent intermetallic compound formation.
Measurement Technique:
- Biasing: A "floating" current source was used to bias the control junction ( $I_{ctrl}$ ) while measuring the differential resistance of the sample junction ( $I_{smpl}$ ). This ensured a single grounding point to minimize noise.
- Feedback Loop (for SQUID): A custom PID controller locked to the resistive peak of the SQUID transition. This allowed real-time tracking of the critical current ( $I_c$ ) as a function of both external magnetic flux and the control current bias.

3. Key Contributions

Non-Dissipative Tuning Mechanism: The paper demonstrates that $I_c$ (and consequently $E_J$ ) of a single SNS junction can be tuned by driving a supercurrent through a separate, coupled junction sharing the same weak link, eliminating the need for a flux loop.
Diffusive Normal Metal Implementation: Unlike previous multiterminal studies that used ballistic graphene, this work utilizes diffusive normal metals (Au), offering greater design flexibility and compatibility with standard fabrication processes.
Characterization of Correlated Transport: The study maps the Critical Current Contours (CCCs) and analyzes how the phase difference induced in one junction affects the current-phase relation (CPR) of the coupled junction.

4. Results

Monotonic Suppression of Critical Current:
- In the 4TJJ geometry, applying a supercurrent bias ( $I_{ctrl}$ ) to the control junction causes a monotonic suppression of the sample junction's critical current ( $I_c$ ).
- Example: Driving 4 $\mu$ A of supercurrent reduced the sample $I_c$ by approximately 50%.
- At higher biases, the Josephson coupling across the sample junction is destroyed, transitioning the device to a resistive state.
SQUID Behavior and Non-Sinusoidal CPR:
- In the SQUID configuration, the total critical current oscillates with external magnetic flux (period $\Phi_0 = h/2e$ ).
- Non-Sinusoidal Oscillations: At $I_{ctrl} = 0$ , the CPR is skewed, indicating higher harmonics. As $I_{ctrl}$ increases, a phase difference ( $\phi_{ctrl}$ ) develops, further modifying the CPR.
- Tunability: The total Josephson energy of the SQUID loop can be reduced by approximately 20% in situ using the supercurrent drive.
- Current Routing: The data suggests that at specific bias points, supercurrent preferentially routes through one branch of the SQUID due to modified quasiparticle density of states, yet coherent $h/2e$ oscillations persist.
High-Bias Features: The differential resistance maps show "centrifugal streaking" features at high bias, attributed to non-local Andreev processes and enhanced supercurrent flow when terminal potentials are commensurate, rather than bound state resonances.

5. Significance

Noise Reduction: This method offers a pathway to realize frequency-tunable qubits with greatly reduced susceptibility to flux noise, as it removes the need for large magnetic flux loops that act as antennas for environmental noise.
Precision Control: It allows for precise, in situ control of the frequency response of a single junction without the complexity of a two-junction loop.
Scalability: By using diffusive normal metals and standard lithography, this approach is potentially more scalable and compatible with existing superconducting circuit architectures than exotic material-based gatemon approaches.
Fundamental Physics: The results provide experimental evidence for correlated Josephson effects and phase synchronization in multiterminal diffusive junctions, supporting theoretical predictions regarding non-local Cooper quartet transport and phase-dependent CPR modification.

In conclusion, Wisne and Chandrasekhar successfully demonstrated that supercurrent biasing is a viable, non-dissipative alternative to magnetic flux tuning for Josephson energy control, paving the way for more robust and noise-resistant superconducting qubits.