A plug-and-play superconducting quantum controller at millikelvin temperatures enables exceeding 99.9% average gate fidelity

Abstract: The development of large-scale superconducting quantum computing requires efficient in-situ control methods that allow high-fidelity operations at millikelvin temperatures. Superconducting circuits based on Josephson junctions offer a promising solution due to their high speed, low power dissipation, and cryogenic nature. Here, we report a superconducting quantum controller that enables direct chip-to-chip interconnection with qubits at 10 mK and high-fidelity, all-digital manipulation. Randomized benchmarking reveals a uniformly high average Clifford fidelity of 99.9% with leakage to high energy levels on the order of $10^{-4}$, and an estimated average gate operation energy of 0.121 fJ, demonstrating the potential to resolve the control bottleneck in superconducting quantum computing.

• The paper demonstrates a plug-and-play superconducting quantum controller operating at millikelvin temperatures with over 99.9% single-qubit gate fidelity.
• It employs SFQ pulse sequencing and robust spectral engineering to reduce leakage (≈10⁻⁴) and minimize decoherence in qubit control.
• The study confirms that low energy dissipation (0.121 fJ per gate) and simplified interconnects enhance scalability for fault-tolerant quantum computing.

Plug-and-Play Superconducting Quantum Controller Enabling >99.9% Average Gate Fidelity

Overview

The paper "A plug-and-play superconducting quantum controller at millikelvin temperatures enables exceeding 99.9% average gate fidelity" (2604.05693) presents a fully integrated, millikelvin-operable superconducting quantum controller that achieves a uniformly high average single-qubit Clifford fidelity of 99.9%, with leakage rates on the order of $10^{-4}$ and extremely low operation energy. The controller supports direct chip-to-chip connections with minimal wiring complexity, enabling scalable high-performance quantum operations. The architecture leverages single-flux-quantum (SFQ) pulse sequences and robust spectral engineering for digital qubit control, effectively suppressing the decoherence channels endemic to cryogenic signal distribution.

Device Architecture and Millikelvin Performance

The controller is founded on superconducting digital logic, generating quantized SFQ voltage pulses for qubit driving. Critical design elements to achieve high fidelity at millikelvin temperatures include:

• Fully passive superconducting bias network: Eliminates static power, suppressing nonequilibrium quasiparticle generation that degrades coherence in the mK environment.
• Low-critical-current SFQ elements: Reduces dynamically generated quasiparticles during switching.
• On-chip spectrum-engineering unit: Minimizes out-of-band noise, optimizes power transfer, and filters extraneous spectral components from the SFQ pulse sequence.

The device is fabricated on a high-resistivity silicon substrate using a multilayer niobium process that is CMOS-compatible, supporting scalable chip production and integration. Cryogenic compatibility is maintained through direct chip-to-qubit connections at 10 mK, enhancing system efficiency and flexibility. The plug-and-play characteristic is realized via modular interconnections (wire bonding, coaxial jumpers), which optimize the electromagnetic environment and minimize parasitic coupling.

High-Fidelity Clifford Gate Benchmarking

The experimental platform achieves coherent qubit control with arbitrary pulse durations, demonstrated through Rabi oscillations and Ramsey interference experiments. Gates of interest ($\mathrm{X}_{\mathrm{SQC}}$, $\mathrm{X}_{\mathrm{SQC}}/2$) are calibrated with narrow distributions of duration and SFQ count.

Randomized benchmarking (RB) of all 24 Clifford gates and interleaved RB (IRB) for primitive gates establish:

• Average Clifford fidelity: 99.90(1)%
• Primitive SFQ $\pi/2$-gate fidelity: 99.90(2)%
• Gate error rates: $1\times10^{-3}$, several times below recent records with similar SFQ-based control hardware

Phase-frame updates using virtual-$\mathrm{Z}$ gates further reduce gate duration and error by minimizing pulse overhead. The reported fidelity distribution is tightly clustered for the pulsed Clifford set, and virtual-$\mathrm{Z}$ operations are idealized via reference frame manipulation.

Suppression of Leakage and Quasiparticle Poisoning

Purity and leakage analyses show:

• Purity RB reports a decoherence-limited error per Clifford of $8.91\times10^{-4}$, essentially matching the RB average error, indicating performance at the coherence boundary.
• Leakage per Clifford: $2.90\times10^{-4}$ with the superconducting controller, $1.46\times10^{-4}$ with conventional microwave drive—indicating leakage is suppressed to a level comparable with room-temperature electronics. Spectral residue in the SFQ drive is identified as a source of the slight discrepancy, addressable by advanced spectrum engineering (pulse shaping, quadrature correction).

The architecture’s fidelity improvements arise from comprehensive suppression of both phonon-mediated and photon-induced quasiparticle poisoning (via passive network and spatial shielding) and high-energy-level leakage (via spectral filtering of SFQ pulses).

Energy Dissipation and Thermal Load

The average energy for a single Clifford operation is 0.121 fJ, orders of magnitude below room-temperature and even some cryogenic-capable classical control electronics. Static dissipation is rendered negligible due to the fully passive bias scheme.

Thermal excitation is evaluated via qubit state tomography after long-duration controller operation. No statistically significant increase in excited-state occupation is observed, confirming that stray heating, photon-induced errors, and quasiparticle-related processes do not measurably impact the qubit, even after protracted control operation.

Implications and Future Directions

This demonstration establishes the practical viability of all-digital, plug-and-play superconducting quantum control for scalable fault-tolerant quantum computers. The measured fidelity well exceeds the surface code threshold, supporting logical qubit implementations and practical QEC.

The architecture’s modularity, SFQ-based signaling, and chip-to-chip flexibility strongly address wiring complexity and thermal challenges that typically constrain dilution-refrigerator-based platforms.

Prospective advances to enhance fidelity and extend the gate set include:

• Dual-pulse interval modulation for envelope control [liu2023single2]: Enabling simultaneous leakage mitigation and unitary error suppression through SFQ drive shaping.
• Pulse sequence optimization with machine learning, genetic algorithms, and discrete search [dalgaard2020global, bastrakova2023genetic]: Further reducing error and enabling short gate times.
• Multiplexing architectures and tunable bandpass components: For large-scale qubit control, including programmable all-microwave two-qubit CZ gates [wang2023single, shirai2023all].

From a theoretical standpoint, the data validate that system coherence limits, rather than control electronics, now represent the practical ceiling on gate performance in such platforms. This shifts the challenge toward further improving qubit coherence and developing robust multiplexed control chains.

Conclusion

The realization of a millikelvin-operable, plug-and-play superconducting quantum controller achieving average Clifford fidelities exceeding 99.9% constitutes a significant advancement for scalable quantum computing architectures (2604.05693). The approach resolves key roadblocks in cryogenic control engineering—specifically, extreme wiring complexity, thermal load, and control-induced leakage—thereby facilitating the path toward practical, large-scale, fault-tolerant superconducting quantum processors. Continued development will likely focus on advanced gate shaping, spectral engineering, and scalable multiplexed integration to further approach the intrinsic limits imposed by qubit coherence and noise.