Skip to content
Quantum.world

Research

Autler-townes energy-level engineering for high-quality all-microwave qubit gate

The gate is calibrated by first preparing the (|11\rangle) state, rapidly increasing the AT drive amplitude to bring the interacting state into resonance, waiting for a variable delay, and then…

Autler-townes energy-level engineering for high-quality all-microwave qubit gate
arXiv
2611.07967
Published
2026-11-05
Authors
Wan-Peng Gao, Xiao-Liang He, Zheng-Qi Niu, Si-Qi Li, Da-Qiang Bao, Kuang Liu, Jian-Qiang You, Jaw-Shen Tsai, Zhen Wang, Zhi-Rong Lin

AI summary

Overview

Research area: Superconducting quantum computing, specifically microwave-driven two-qubit entangling gates and energy-level engineering. Technical level: Intermediate — accessible to readers with basic familiarity with qubits, transmon anharmonicity, and randomized benchmarking. Scope: The paper proposes and experimentally validates an Autler-Townes-based energy-level engineering technique that enables a fast, high-fidelity all-microwave CZ gate in coupled fixed-frequency transmon qubits.

What This Paper Is About

High-fidelity entangling gates are required for fault-tolerant superconducting quantum computing, but existing approaches face a trade-off: flux-tuning methods compromise coherence and flux-noise immunity, while all-microwave schemes trade gate speed against leakage. The paper aims to tune qubit nonlinearity using the Autler-Townes effect without perturbing the computational subspace or degrading coherence time, and thereby realize a fast all-microwave CZ gate.

Key Contributions

  1. Autler-Townes energy-level engineering: A microwave drive near the (|2\rangle \leftrightarrow |3\rangle) transition of one qubit splits the excited levels into Autler-Townes doublets, making the effective anharmonicity tunable while leaving the computational subspace largely unaffected because the drive is far detuned from the (|0\rangle \leftrightarrow |1\rangle) transition.
  2. Experimental demonstration of a fast CZ gate: Using two coupled Xmon qubits, the authors demonstrate a 97 ns CZ gate with a fidelity of 99.40(9)%, extracted from reference and interleaved randomized benchmarking.
  3. Error-budget analysis: The gate error is dominated by static decoherence, estimated at 0.0047, or approximately 78% of the total error; the remaining approximately 22% is mostly attributed to an average parasitic X-rotation angle of 0.08 rad, corresponding to a gate infidelity of 0.0013.
  4. Proposed AT coupler scheme: The technique is extended to a three-step scheme using an AT coupler that cancels residual coupling between data qubits and is compatible with prevailing coupler structures; the same mechanism could also generate entanglement between a second qubit and a third qubit, equivalent to a Toffoli gate with single-qubit Z rotations.

Main Findings

  • Gate fidelity and speed: The CZ gate operates in 97 ns and reaches (F_{\mathrm{CZ}} = 0.9940(9)). Randomized benchmarking gives (p_{\mathrm{ref}} = 0.9575(7)) for the reference sequence and (p_{\mathrm{int}} = 0.9498(9)) for the interleaved sequence, yielding an error rate (r_{\mathrm{CZ}} = 0.0060(9)). The fidelity surpasses the surface code error-correction threshold.
  • Tunable anharmonicity observed spectroscopically: Two-tone spectroscopy with the AT drive set to 4.904 GHz, near the (|2\rangle \leftrightarrow |3\rangle) transition of (Q_1), shows that the AT doublet splitting increases with drive amplitude (\Omega_{\mathrm{AT}}). With (Q_2) fixed at 5.124 GHz, an avoided crossing appears between the upper branch and (Q_2), indicating transverse coupling.
  • Effective coupling strength: The measured effective coupling is (g_{\mathrm{mea}}/2\pi \simeq 5.9) MHz, supporting fast gate operation. Theory gives (\sqrt{2}g \lesssim \tilde{g} \lesssim \sqrt{3}g), depending on the mixing angle of the AT doublet.
  • Coherence preserved: The coherence times (T_1) and (T_2^e) of (Q_1) remain almost unaffected by varying AT drive amplitude. Device parameters include (Q_1) and (Q_2) frequencies of approximately 5.272 GHz and 5.124 GHz, anharmonicities of approximately -178 MHz and -205 MHz, (T_1) of 27 µs and 17 µs, and (T_2^e) of 36 µs and 22 µs.
  • Small frequency drift: A frequency drift of less than about 2 MHz around the CZ working point is observed and attributed to off-resonant driving of (Q_1)'s (|1\rangle \leftrightarrow |2\rangle) transition. Parasitic phase accumulation is compensated with virtual-Z operations on (Q_1) and (Q_2).
  • Error dominated by static decoherence: Using (r_{\mathrm{relaxation}} = \frac{2}{5}\tau_{\mathrm{gate}}/T_1) and (r_{\mathrm{dephasing}} = \frac{2}{5}\tau_{\mathrm{gate}}/T_\phi), the decoherence contribution is estimated at 0.0047, approximately 78% of the total gate error, indicating that improved coherence time would directly improve gate fidelity.
  • Enhanced coupler concept: In the proposed AT coupler scheme, a three-step pulse sequence brings (Q_1) and the lower branch of the coupler into resonance, then (Q_2) and the upper branch into resonance, and finally returns (Q_1) and the lower branch into resonance. The process is equivalent to a CZ gate with single-qubit Z rotations.

Methodology in Plain English

The researchers fabricated two coupled Xmon qubits from a 100 nm thick tantalum/niobium bilayer on a sapphire (111) substrate, with aluminum air-bridges over each coplanar waveguide, and cooled the device in a BlueFors XLD-1000 dilution refrigerator at a base temperature of about 10 mK. Rather than changing a qubit's frequency with a flux pulse, they apply a strong microwave tone close to the (|2\rangle \leftrightarrow |3\rangle) transition of one qubit. This drive dresses the higher energy levels and splits them into Autler-Townes doublets; the split depends on the drive amplitude, so the qubit's effective anharmonicity becomes adjustable. Because the drive is far from the (|0\rangle \leftrightarrow |1\rangle) transition and the qubit has large anharmonicity, the computational states are mostly left alone.

The gate is calibrated by first preparing the (|11\rangle) state, rapidly increasing the AT drive amplitude to bring the interacting state into resonance, waiting for a variable delay, and then ramping the drive down. Measuring populations versus drive amplitude and pulse length produces Chevron patterns that reveal coherent population exchange and allow the 97 ns operation point to be selected. Gate fidelity is then measured with Clifford-based randomized benchmarking: 300 random sequences per sequence length, with reference and interleaved sequences, and error bars representing 95% confidence. An error budget is built from measured (T_1) and pure dephasing times plus a numerical simulation of an off-resonantly driven qutrit.

Why This Matters

The work shows that the Autler-Townes effect can be used as a practical knob for quantum-level engineering in superconducting circuits, avoiding flux tuning and its associated noise and coherence penalties. The resulting gate is all-microwave, fast, and compatible with fixed-frequency qubit architectures, which are attractive for scaling because they avoid flux-noise sensitivity and simplify control wiring.

Real-world applications:

  • Fault-tolerant quantum processors: A 99.40(9)% CZ gate above the surface-code threshold is a building block for error-corrected logical qubits.
  • Quantum algorithm hardware mapping: The proposed AT coupler could implement Toffoli-like operations with fewer two-qubit gates than a standard decomposition, reducing overhead in compiled quantum circuits.
  • Quantum simulation and chemistry: Faster, high-fidelity entangling gates improve circuit depth and fidelity for simulating molecular and material systems.
  • Multi-qubit superconducting hardware: The scheme is compatible with prevailing coupler structures and could be integrated into larger fixed-frequency processor layouts.

Industry relevance: The approach uses standard thin-film superconducting fabrication, conventional microwave drive electronics, and fixed-frequency qubit designs, making it relevant to foundry-style quantum processor manufacturing and to companies building microwave-controlled superconducting quantum computers.

Future Directions

  • Improve coherence times: Since static decoherence accounts for approximately 78% of the gate error, materials and device improvements that extend (T_1) and (T_\phi) are the most direct route to higher fidelity.
  • Optimize pulse shapes: The remaining approximately 22% of the error is mostly from parasitic X rotations; shaped or optimized AT pulses could suppress this control error.
  • Implement the AT coupler in larger arrays: The three-step coupler scheme and the proposed Toffoli-like operation need experimental validation in multi-qubit processors with residual-coupling cancellation.
  • Extend to other platforms: The authors argue the technique is general and could be applied to other quantum platforms that require energy-level engineering.

Target Audience

This paper benefits superconducting-qubit experimentalists, quantum-control engineers, graduate students working on two-qubit gates, and researchers in quantum error correction who need to understand the fidelity and error budget of a microwave-only CZ gate. Readers with an intermediate background in quantum information will gain the most, while specialists will appreciate the device parameters, randomized-benchmarking data, and error analysis.

Authors’ abstract

High-fidelity entangling gates are essential for realizing large-scale fault-tolerant superconducting quantum computing, yet existing approaches are limited: frequency-tuning techniques induce decoherence, while all-microwave-driven schemes suffer the trade-off between gate speed and leakage rate. Here, we introduce an energy-level engineering technique that employs the Autler-Townes effect to manipulate the nonlinearity of qubits without perturbing the computational…

The full abstract on arXiv
Read the original paper