[論文レビュー] Mid-circuit measurements on a single species neutral alkali atom quantum processor

T. M. Graham, L. Phuttitarn+8 more|arXiv (Cornell University)|Mar 17, 2023

Quantum Information and Cryptography参考文献 56被引用数 15

ひとこと要約

保護されたハイパーファイン状態にデータ量子ビットを棚上げしつつ、非破壊的に ancilla を測定することで Cs 中性原子アレイにおけるスケーラブルな中間回路測定を実証。SPAM補正忠実度を高く達成し、データ量子ビットのコヒーレンスを保持。

ABSTRACT

We demonstrate mid-circuit measurements in a neutral atom array by shelving data qubits in protected hyperfine-Zeeman sub-states while non-destructively measuring an ancilla qubit. Measurement fidelity was enhanced using microwave repumping of the ancilla during the measurement. The coherence of the shelved data qubits was extended during the ancilla readout with dynamical decoupling pulses, after which the data qubits are returned to mf = 0 computational basis states. We demonstrate that the quantum state of the data qubits is well preserved up to a constant phase shift with a state preparation and measurement (SPAM) corrected process fidelity of F = 97.0(5)%. The measurement fidelity on the ancilla qubit after correction for state preparation errors is F = 94.9(8)% and F = 95.3(1.1)% for |0> and |1> qubit states, respectively. We discuss extending this technique to repetitive quantum error correction using quadrupole recooling and microwave-based quantum state resetting.

研究の動機と目的

• Demonstrate mid-circuit measurement in a single-species neutral Cs atom processor.
• Protect data qubits by shelving into hyperfine states during ancilla readout.
• Quantify ancilla readout fidelity and the impact on data-qubit coherence.
• Show SPAM-corrected process fidelity approaching fault-tolerant-relevant levels and discuss paths to qubit reset and error correction.

提案手法

• Shelving data qubits into f=3 hyperfine manifold to decouple from readout light.
• Non-destructive, state-selective readout of ancilla using fluorescence detection.
• Dynamical decoupling (three-pulse echo sequence) applied to data qubits during ancilla readout.
• Microwave control with two horns to selectively drive shelving transitions and manage polarization.
• Ancilla readout uses detuned, repumped light with site-selective Stark shifts to protect the ancilla.
• SPAM-corrected process fidelity quantified via Ramsey-type benchmarking of data qubits and ancilla readout fidelity assessment.

Figure 1: a) Experimental layout for mid-circuit measurement. Atoms were trapped and cooled into a $3\times 3$ 1064-nm tweezer array. The microwave horns are each attached to 40 W amplifiers that are driven by 9.2 GHz signals. A 459-nm beam provided site-selective Stark shifts on targeted sites. Dur

実験結果

リサーチクエスチョン

• RQ1 Can mid-circuit measurements be implemented in a single-species neutral-atom array without destroying data-qubit states?
• RQ2 What are the achievable fidelities for ancilla readout and for preserving data-qubit coherence during the mid-circuit measurement?
• RQ3 How does shelving data qubits and performing echo decoupling affect the overall quantum process fidelity of the data qubits and ancilla?
• RQ4 What improvements (e.g., qubit resetting, faster rotations, recooling) are needed to reach fault-tolerant operation?

主な発見

• Ancilla readout fidelity (after SPAM correction) ≈ 97.0% for data-qubit preservation under mid-circuit measurement.
• Ancilla measurement fidelities are ≈ 94.9% for |0> and 95.3% for |1> after SPAM correction.
• Raw data-qubit process fidelity during mid-circuit measurement ≈ 93.8%; SPAM-corrected process fidelity ≈ 97.0%.
• Data qubits retain coherence during the ancilla readout with measured Ramsey-based phase shift consistent with a unitary phase gate (average process fidelity ~93.8% raw, 97.0% SPAM-corrected).
• Coherence preservation is enabled by dynamical decoupling (8 Hahn echo pulses) and by carefully balancing readout detuning to limit heating and dephasing.
• Identified pathways to scalable qubit reset and repetitive error correction via quadrupole recooling and microwave-based resetting.

Figure 2: Mid-circuit measurement results. a) To show that the data qubits retain coherence during the mid-circuit measurement, we prepared the atoms in $\frac{1}{\sqrt{2}}\left(\ket{0}+\ket{1}\right)$ using a $\pi/2$ -microwave pulse and then performed a mid-circuit measurement on the ancilla. Afte