⚛️ quantum physics

Electrical post-fabrication tuning of aluminum Josephson junctions at room temperature

This paper demonstrates a room-temperature electrical tuning method for aluminum Josephson junctions using voltage pulses to controllably increase resistance and adjust qubit frequencies by up to 2 GHz while maintaining quality factors above 1 million, offering a practical solution for post-fabrication mitigation of frequency crowding in superconducting quantum processors.

Original authors: Christian Križan, Maurizio Toselli, Irshad Ahmad, Hadi Khaksaran, Marcus Rommel, Nermin Trnjanin, Janka Biznárová, Mamta Dahiya, Emil Hogedal, Halldór Jakobsson, Andreas Nylander, Jonas Bylander, Per

Published 2026-06-23

📖 4 min read🧠 Deep dive

CC BY 4.0

Original authors: Christian Križan, Maurizio Toselli, Irshad Ahmad, Hadi Khaksaran, Marcus Rommel, Nermin Trnjanin, Janka Biznárová, Mamta Dahiya, Emil Hogedal, Halldór Jakobsson, Andreas Nylander, Jonas Bylander, Per Delsing, Giovanna Tancredi

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

Imagine a supercomputer built not from silicon chips, but from tiny loops of wire that conduct electricity with zero resistance. These are superconducting quantum computers. The heart of these machines is a microscopic component called a Josephson junction. Think of a junction as a very specific type of "traffic light" for electricity. It controls how fast the computer's "bits" (qubits) can think.

For the computer to work, every single traffic light needs to be set to the exact same speed. However, when scientists build these chips, tiny imperfections in the manufacturing process mean that some traffic lights are slightly faster or slower than others. This is like trying to tune a choir where everyone is singing slightly off-key; eventually, their voices clash, and the music (the computer's calculations) becomes a mess. This problem is called "frequency crowding."

Usually, once these chips are built, you can't fix the traffic lights. If they are out of tune, the chip is often useless.

The New "Tuning Knob"

This paper presents a new way to fix these traffic lights after the chip is built, using nothing but electricity at room temperature.

The researchers discovered that if you zap a junction with specific voltage pulses (like quick, sharp taps of electricity), you can permanently change its resistance. Think of this resistance as the "tightness" of a spring. By applying these electrical taps, they can loosen or tighten the spring to make the traffic light run at the exact speed needed.

Here is how they did it and what they found:

1. The "Tap" Method
Instead of heating the whole chip (which would be like trying to fix one violin string by heating the entire orchestra), they used a precise sequence of electrical pulses. They applied a positive voltage, then a negative one, over and over again.

The Result: The more they tapped (and the harder they tapped), the more the resistance changed. They found that a small increase in the "strength" of the tap caused a massive, exponential increase in the change of resistance. It's like turning a volume knob where a tiny twist makes the sound get much, much louder.

2. The "Settling" Effect
There was a catch. When they stopped tapping, the junction didn't stay exactly where they left it. It kept changing on its own for a while, like a swing that keeps moving back and forth after you stop pushing it.

The Discovery: They found that this "settling" behavior was predictable. The more they changed the resistance on purpose, the more it would change on its own afterward. They had to wait for the junction to "calm down" before they knew the final result.

3. The "Step-by-Step" Strategy
At first, they tried to change the resistance all at once, but the junctions would sometimes break (like snapping a rubber band by pulling too hard).

The Solution: They developed a "stepped" approach. They would tap the junction a little bit, wait for it to settle, tap it a little more, wait again, and repeat.
The Achievement: Using this careful, step-by-step method, they were able to increase the resistance of a junction by 270%. This is a huge change, allowing them to shift the frequency of a qubit by nearly 2 GHz (gigahertz). This is enough to fix the "frequency crowding" problem and save a quantum processor from being unusable.

4. Did it break the computer?
The biggest worry was: "Does zapping these delicate components ruin the computer's ability to think?"

The Good News: They tested this on actual qubits. Even after tuning them, the "quality" of the qubits remained incredibly high (over 1 million). It's like tuning a piano string so precisely that the note is perfect, but the string itself is still just as strong and durable as before.

Summary

In simple terms, this paper shows that we can take a quantum computer chip that was built with slightly "out-of-tune" parts and fix it using electrical pulses at room temperature. By tapping the parts with electricity in a specific, step-by-step rhythm, they can adjust the parts to the perfect speed without breaking them. This offers a way to rescue expensive quantum processors that would otherwise be discarded due to tiny manufacturing errors.

Technical Summary: Electrical Post-Fabrication Tuning of Aluminum Josephson Junctions at Room Temperature

Problem Statement
Superconducting quantum processors face significant challenges in fabrication precision, particularly regarding the spread of Josephson junction resistances. Even with state-of-the-art lithography and deposition techniques, resistance variations on the order of 2% across a wafer translate to frequency spreads of approximately ±50 MHz for transmon qubits. This "frequency crowding" leads to spectral collisions between quantum operations, reducing the probability of collision-free operations to single-digit percentages for processors with as few as 30 qubits. Scaling to 1,000 qubits demands frequency placement tolerances of only a few MHz. While post-fabrication tuning via thermal annealing or laser heating exists, these methods often lack the selectivity to tune individual junctions or require complex engineering of the surrounding electromagnetic environment. Electrical tuning, which modifies junction resistance via applied voltage pulses, offers a promising alternative but requires a deeper understanding of its mechanisms, limits, and relaxation behaviors to be viable for large-scale processors.

Methodology
The authors investigated electrical tuning on Al-AlOx-Al Josephson junctions fabricated on intrinsic Si substrates. The study utilized three wafers with varying oxidation doses (low, medium, and high) and junction sizes (electrode widths from 150 nm to 600 nm).

The core experimental procedure involved active resistance manipulation using a bipolar voltage pulse train. This sequence consisted of a positive-bias square pulse, a blanking interval, a negative-bias square pulse, and a second blanking interval. The standard protocol applied pulses with an amplitude of ±0.85 V and a duration of 200 ms, separated by 100 ms intervals. Resistance was monitored via four-wire I-V measurements at room temperature.

Key experimental variations included:

Aging Characterization: Monitoring spontaneous resistance increases over time under ambient conditions to determine if measurement cadence or current influenced aging.
Voltage Dependence: Testing pulse amplitudes to model the rate of resistance change.
Relaxation Analysis: Observing the spontaneous resistance increase immediately following the cessation of active manipulation, measured over periods up to 24 hours.
Stepped Manipulation: Implementing a protocol where active manipulation was paused after a fixed resistance increment to allow for relaxation, then resumed, to extend the total achievable resistance change.
Qubit Performance: Fabricating two devices (S1 and S2) containing eight transmon qubits each. These were subjected to one or two rounds of manipulation, followed by cryogenic characterization of energy relaxation time ( $T_1$ ), Ramsey dephasing time ( $T_2^*$ ), and Hahn-echo dephasing time ( $T_2^e$ ) to calculate quality factors ( $Q$ ).

Key Results

Aging and Measurement Impact: Natural aging causes junction resistance to increase spontaneously by several tens of percent over a month. Smaller junctions exhibit faster aging. Crucially, the study found that routine resistance measurements (ranging from daily to bimonthly) do not measurably affect the aging process.
Active Manipulation Dynamics: The rate of resistance increase during active manipulation scales exponentially with the applied pulse amplitude ( $V$ ). The linear coefficient of resistance change, $\alpha(V)$ , follows the relation $\alpha(V) = \alpha_0 \cdot e^{V/V_0}$ , where the characteristic voltage $V_0$ ranges from 55 to 91 mV. An increase of approximately 200 mV in pulse amplitude can result in an order-of-magnitude increase in the tuning rate for low- and high-dose junctions.
Relaxation Behavior: Upon stopping active manipulation, the resistance continues to increase spontaneously. This relaxation is not merely a fixed offset; it scales linearly with the amount of resistance change induced during active manipulation. The relaxation process exhibits a logarithmic time dependence. The study determined that the relaxation contributes an additional resistance change of approximately 3.29% independent of active manipulation, plus a component proportional to the active change.
Stepped Protocol: By employing a stepped manipulation protocol (manipulate, relax, repeat), the authors achieved a total resistance increase of 269.5% (nearly 270%) in a single junction. This corresponds to a transmon qubit frequency shift of approximately -1964 MHz. This represents the largest targeted resistance manipulation reported to date.
Qubit Quality: Despite the manipulation, qubit quality factors remained preserved above 1 million (specifically between 1 and 1.5 million) after a single round of tuning, comparable to standard fabrication outcomes. No systematic dependence of quality factors on the manipulation process was observed, although some junctions failed during the process, likely due to fabrication defects in the specific wafer used rather than the tuning protocol itself.

Significance and Claims
The paper establishes that room-temperature electrical tuning of aluminum Josephson junctions is a feasible method for post-fabrication correction of qubit frequencies. The authors demonstrate that:

Tunability: Junction resistance can be controllably increased by up to 270% using electrical pulses, offering a solution to frequency crowding in quantum processors.
Predictability: The tuning rate is exponentially dependent on voltage amplitude, and the subsequent relaxation behavior is quantifiable, scaling with both the initial resistance and the magnitude of the active manipulation.
Non-Destructive Nature: When successful, the tuning process does not degrade qubit coherence or quality factors, maintaining performance levels consistent with standard fabrication.
Practical Limits: The study identifies the relaxation behavior as a critical factor for precision tuning. The uncertainty in the relaxation process sets a lower bound on achievable frequency precision (estimated at 11.3 MHz), which is significantly better than standard fabrication precision but requires multi-day monitoring of sacrificial junctions to characterize accurately.

The work provides a practical framework for mitigating frequency crowding, defining the achievable range, relaxation behavior, and practical limits of electrical tuning for superconducting quantum processors.

The New "Tuning Knob"

Summary

Technical Summary: Electrical Post-Fabrication Tuning of Aluminum Josephson Junctions at Room Temperature

More like this