Revisiting the multi-mode rhombus circuit as a biased-noise qubit

This paper revisits the multi-mode rhombus circuit as a biased-noise qubit by intentionally altering a junction's energy to enable direct GHz-range probing, demonstrating that operating away from half-flux frustration yields significantly improved relaxation times ($T_1 \approx 500\,\mu$s) compared to the frustrated regime, with loss analysis identifying flux noise and quasiparticle tunneling as key limiting factors.

The Gist

Imagine you are trying to build a super-precise digital switch (a qubit) that can hold a secret without the outside world messing it up. In the world of quantum computing, the biggest enemy is "noise"—tiny, random jitters from the environment that cause the switch to flip its state or lose its memory.

For a long time, scientists have tried to build "protected" switches that are naturally immune to these jitters. One famous design is called the Rhombus Qubit. Think of this like a perfectly balanced seesaw with four wheels. If you set it up just right (with a specific magnetic field), the two sides of the seesaw are so perfectly symmetrical that a tiny nudge from the left is exactly canceled out by a nudge from the right. Theoretically, this makes it impossible for the switch to accidentally flip its state due to electrical noise.

The Problem with the Perfect Seesaw
However, the original Rhombus design had a flaw. While it was great at ignoring electrical jitters, it was very sensitive to magnetic jitters and tiny particles called "quasiparticles" (which are like broken pieces of the superconducting material). It was like building a boat that was waterproof but had a hole in the bottom; it could handle rain (electrical noise) but would sink if a wave (magnetic noise) hit it. Also, the original design operated at a very low frequency, which made it even more vulnerable to these magnetic waves.

The New Idea: The "Soft" Rhombus
In this paper, the researchers decided to break the perfect symmetry on purpose. They intentionally made one of the four wheels on their seesaw slightly smaller (less energetic) than the others. They call this the "Soft-Rhombus Qubit."

Here is why this "imperfect" design is actually better:

1. Raising the Frequency: By making the wheel smaller, they raised the "pitch" of the seesaw. Instead of a low, slow hum, it now vibrates at a higher, faster frequency.
2. Avoiding the Noise: The main sources of noise (magnetic jitters and quasiparticles) are strongest at low frequencies. By moving the qubit to a higher frequency (around a few GHz), they effectively moved the switch out of the "loud" part of the noise spectrum.
3. The Biased-Noise Trade-off: This change creates a new type of protection. The qubit is no longer protected against all errors equally. Instead, it becomes a "biased-noise" qubit. This means it is very good at resisting one type of error (relaxation, or losing energy) but slightly more vulnerable to another (dephasing, or losing its timing).

The Experiment
The team built this new circuit using standard materials (aluminum and tantalum) on a sapphire chip. They tested it by measuring how long the qubit could hold its state before failing.

• At the "Frustrated" Point (The old way): When they used the magnetic field to make the qubit perfectly balanced (like the original design), it was very sensitive to magnetic noise. It lost its energy quickly (in about 27 microseconds) and its timing got messy quickly.
• At the "Biased" Point (The new way): When they moved the magnetic field slightly away from that perfect balance, the qubit's behavior changed. It became much more stable against losing energy. They measured a relaxation time of about 500 microseconds (nearly 20 times longer than before!).

The Conclusion
The paper concludes that while the "perfect" symmetric design sounds great on paper, it fails in the real world because of magnetic noise and quasiparticles. By intentionally making the circuit "soft" and asymmetric, they created a qubit that is much more robust against the specific types of noise that actually exist in a lab.

They found that there is a "sweet spot" operating frequency (a few GHz) where this qubit works best. In this regime, the qubit acts like a very durable container that holds its energy for a long time, even though it might get its timing slightly scrambled. This suggests that for building future quantum computers, it might be better to design circuits that are "imperfect" in a specific way to fight real-world noise, rather than trying to be perfectly symmetric.

Technical Summary

Technical Summary: Revisiting the Multi-Mode Rhombus Circuit as a Biased-Noise Qubit

Problem Statement
The rhombus qubit, originally proposed as a building block for topologically protected array qubits, relies on interference between pairs of Josephson junctions to encode information in charge-parity states. While the ideal rhombus circuit offers protection against dielectric-type relaxation processes (local charge noise), it suffers from significant vulnerabilities in realistic environments. Specifically, the interferometer-based layout and its typically low transition frequencies expose the qubit to flux noise and quasiparticle tunneling. In the original proposal, protection is strictly maintained only at exactly half a flux quantum ($\Phi_{ext} = \Phi_0/2$); deviations from this "frustration" point due to flux noise cause the protection to vanish, rendering the qubit energy linearly sensitive to magnetic fields. Furthermore, the low operating frequencies of ideal rhombus qubits place them in a regime where $1/f$ flux noise and quasiparticle effects are most detrimental to relaxation times ($T_1$).

Methodology
The authors revisit the rhombus circuit by intentionally introducing asymmetry to create a "soft-rhombus" qubit. Instead of maintaining four identical junctions, they lower the Josephson energy of one junction relative to the others (by a ratio $\alpha \approx 0.63$). This modification serves two primary purposes:

1. Frequency Elevation: It lifts the quasi-degeneracy of the qubit states at frustration, raising the transition frequency to the hundreds of MHz to low GHz range.
2. Noise Engineering: It tailors the protection mechanism away from ideal charge-parity symmetry toward a "biased-noise" regime.

The study employs a full multi-mode circuit quantization approach, modeling the device with three degrees of freedom and a Hamiltonian that includes self- and cross-charging energies, as well as the Josephson energies of the four junctions. Theoretical predictions for wavefunctions, energy spectra, and noise sensitivity (charge, flux, and quasiparticle) are compared against experimental data from a device fabricated on a sapphire substrate with tantalum capacitor pads and double-angle evaporated aluminum-oxide Josephson junctions. The device is measured using standard resonator spectroscopy, Ramsey, and echo sequences in a dilution refrigerator.

Key Contributions and Results

• Soft-Rhombus Design: The authors demonstrate that intentionally breaking the symmetry of the junctions allows the qubit to operate at higher frequencies (up to $\sim$1 GHz) while retaining charge insensitivity due to large shunting capacitors. This design shifts the operating point away from the strict $\Phi_0/2$ sweet spot required for perfect charge-parity protection.
• Biased-Noise Regime: Away from flux frustration, the qubit wavefunctions become localized in disjoint phase valleys. This localization leads to a "biased-noise" regime where bit-flip errors (relaxation) are exponentially suppressed, while phase-flip errors (dephasing) remain dominant.
• Experimental Performance:
  • In the biased-noise regime (away from frustration, $\sim$1 GHz), the device achieves an average relaxation time $T_1 \approx 500$ $\mu$s, with a Ramsey dephasing time $T_\phi^R \approx 90$ ns.
  • At frustration (low frequency, $\sim$100 MHz), where protection is theoretically maximal against charge noise, the relaxation time drops significantly to $T_1 \approx 27$ $\mu$s, while dephasing improves to $T_\phi^R \approx 670$ ns.
• Noise Analysis: The authors identify that at low frequencies, the relaxation times are limited primarily by flux noise and quasiparticle tunneling, rather than charge noise. Theoretical models using a flux noise amplitude of $A_\Phi = 4$ $\mu\Phi_0$ and a quasiparticle density of $x_{qp} = 10^{-8}$ align well with the observed trends, confirming that the low-frequency operation is hindered by these specific environmental couplings.

Significance
The paper argues that for certain protected qubits, strict adherence to ideal symmetry (like the perfect rhombus) may not be optimal in the presence of realistic noise sources. By compromising the perfect charge-parity protection to raise the operating frequency and reduce sensitivity to flux and quasiparticle noise, the soft-rhombus qubit achieves significantly longer relaxation times in a biased-noise regime. The authors conclude that low-frequency noise is a critical limiting factor for protected qubits and that design compromises based on realistic noise spectra are essential for optimizing performance. This work suggests that operating in a biased-noise regime, potentially paired with tailored error correction codes, offers a viable path forward for protected superconducting qubits.