Qutrit entanglement and joint multi-parameter estimation in an optical clock platform

AI summary

Overview

• Research area: Quantum metrology and quantum information science with neutral atoms — specifically, high-dimensional (qutrit) entanglement and multi-parameter estimation in an optical clock platform.
• Technical level: Advanced. The paper assumes familiarity with entanglement certification (Schmidt number, fidelity bounds), the quantum Cramér-Rao bound, multi-parameter estimation theory, and neutral-atom quantum control (Rydberg gates, optical tweezers, magic trapping).
• Scope: This summary covers the paper's experimental demonstration of genuine two-qutrit entanglement between ⁸⁸Sr atoms in triple-magic tweezers, its use for joint two-parameter phase estimation across two clock transitions, and the accompanying theoretical analysis of noise-robust scaling.

What This Paper Is About

Quantum-enhanced metrology has almost always used two-level qubits to estimate one parameter at a time. Encoding information in three or more levels (qudits) could let a single entangled probe measure several parameters simultaneously — but genuine high-dimensional entanglement between individual neutral atoms, stored in their internal energy levels, had not been demonstrated, and joint multi-parameter estimation had not been performed with individually controlled atoms.

This paper closes that gap: the authors generate a maximally entangled two-qutrit state from pairs of ⁸⁸Sr atoms with fidelity F = 0.85(1), certifying genuine qutrit entanglement, and then use such states to jointly estimate two phases imprinted on two optical clock transitions.

Key Contributions

1. Genuine qutrit entanglement between individual neutral atoms. A maximally entangled two-qutrit state is prepared using global qutrit control plus a Rydberg-based controlled-Z gate, and certified with a loss-postselected fidelity of F = 0.85(1), exceeding the 2/3 threshold that rules out Schmidt number ≤ 2.

2. An optimal two-qutrit metrological probe and noise-robust readout circuit. The authors theoretically identify the optimal state for joint estimation of two commuting phases and design a readout circuit, tailored to which gates perform best in their apparatus, that saturates the quantum Cramér-Rao bound at Y_joint = 0.4812 for the two-qutrit case.

3. Experimental joint two-parameter estimation below the individual-estimation bound. Using a 29,594-shot dataset and maximum-likelihood estimation via bootstrapping, they measure ν[(Δθ₁)² + (Δθ₂)²] = 0.82(3) < 1, beating the bound achievable by ideal individual estimation with qubit entanglement.

4. A scalable circuit family and noise analysis. Beyond two atoms, they exhibit a family of state-preparation and readout circuits (Fig. 4) and show theoretically that the joint-estimation advantage survives circuit-level depolarizing and dephasing noise at state-of-the-art entangled-atom numbers.

Main Findings

• Fidelity and certification: F = 1/3(C₀₁ + C₀₂ + C₁₂ + P₀₀ + P₁₁ + P₂₂) = 0.85(1) > 2/3. The measured parity contrasts are C₁₂ = 0.56(1), C₀₁ = 0.51(1), and C₀₂ = 0.52(2), against an ideal contrast of 2/3.

• Loss handling: The fidelity is not corrected for state-preparation and measurement errors, apart from discarding events in which one or both atoms of a pair were lost — between 15% and 20% across the different datasets.

• Optimal probe state is unbalanced: Theoretically, the optimal N-atom state is √0.38|0⟩^⊗N + √0.24|1⟩^⊗N + √0.38|2⟩^⊗N. The imbalance reflects the fact that |1⟩ is acted on by both generators. This state is genuinely qutrit-entangled and outperforms individual estimation for any number of atoms.

• Theoretical joint-estimation advantage: Y_joint ≥ (11 + 2√10)/(9N²) ≈ 0.48 × Y_ind, where Y_ind ≥ 4/(νN²) is the bound for ideal individual estimation using a split array with qubit entanglement.

• Measured joint variance: ν[(Δθ₁)² + (Δθ₂)²] = 0.82(3) < 1 — below the ideal individual-estimation threshold, even though the measured value includes experimental imperfections while the bound assumes perfect individual estimation.

• Gap to the ideal bound explained by gate noise: Independently characterized incoherent damping of the R₀₁ and R₁₂ rotations gives (η_x, η_y, η_z) = (0.9809, 0.9341, 1.0000) for R₀₁ and (0.9868, 0.9420, 1.0000) for R₁₂. A circuit-level simulation using these six parameters plus ideal CZ gates and only two free phase shifts reproduces the data well.

• Readout choice matters: A readout circuit using only single-qutrit gates would reach Y_joint = 0.5178 but is more sensitive to gate errors; the implemented noise-robust sequence instead uses two R₁₂ rotations, a CZ gate, and one R₀₁ gate.

• Noise scaling: Without noise, the joint-estimation advantage is constant for arbitrary atom number; with circuit-level depolarizing noise on rotations and dephasing on individual CZ gates, the advantage persists only up to a finite atom number, but remains favorable for state-of-the-art entangled-state sizes with foreseeable gate fidelities.

Methodology in Plain English

Encoding. Each atom stores a qutrit in three internal states: the ground state ¹S₀ (labeled |0⟩), and two fine-structure clock states ³P₀ (|1⟩) and ³P₂, m_J = 0 (|2⟩). Twelve pairs of ⁸⁸Sr atoms sit in 813 nm optical tweezers. By angle-tuning the magnetic field relative to the tweezer polarization, the trap is made "triple-magic," so all three levels stay coherent.

Control. Arbitrary single-qutrit rotations come in two flavors: R₁₂ rotations between the two clock states, driven by Raman beams, and R₀₁ rotations between the ground state and ³P₀, driven by a three-photon scheme. Entanglement comes from a controlled-Z gate implemented with a single-photon transition from ³P₀ to a Rydberg state (n = 47, ³S₁, m_J = −1).

Readout. The three levels are measured one at a time: the ground-state population is destructively imaged on the ¹P₁ transition at the 1% infidelity level; ³P₂ population is then pumped down with a selective 496 nm repumper and imaged the same way; finally the remaining ³P₀ population is pumped down and measured with a slow lossless image. An atom with no counts in any of the three images is treated as lost.

Entanglement benchmark. The authors prepare (|00⟩ + |11⟩ + |22⟩)/√3 with a sequence of global rotations and CZ gates, then measure diagonal populations and parity oscillations on all three pairwise transitions by scanning the phase of a final π/2 analysis pulse. The contrasts give the off-diagonal magnitudes, and these combine into the fidelity bound.

Joint estimation. Two phases θ₁ and θ₂ are imprinted on the |0⟩–|1⟩ and |1⟩–|2⟩ transitions, described by the commuting generators g₁ = (|0⟩⟨0| − |1⟩⟨1|)/√2 and g₂ = (|1⟩⟨1| − |2⟩⟨2|)/√2. Because the generators commute, the quantum Cramér-Rao bound can in principle be saturated. The experiment prepares the unbalanced optimal two-qutrit state, applies the readout circuit, scans the readout pulse phases to build a probability model over the nine measurement outcomes, fits it with a cosine-sine form, and then extracts maximum-likelihood estimators and their covariance via bootstrapping.

Scaling study. For large N the brute-force circuit optimization becomes intractable, so the authors instead design a scalable circuit family using global single-qutrit rotations and an entangling block C_N = ∏{i<j} CZ{ij}, and simulate its performance under depolarizing noise on rotations and dephasing on each CZ.

Why This Matters

Research impact. This is the first demonstration of genuine high-dimensional entanglement between individually controlled neutral atoms encoded in internal energy levels — a building block that had been missing despite interest in qudit encoding for neutral-atom platforms. It also moves multi-parameter clock metrology from atomic ensembles to individually resolved, individually controlled atoms, and shows that qutrit encoding lets all N atoms address both parameters simultaneously rather than splitting the array in half.

Real-world applications:

• Optical lattice and tweezer clocks: Multi-parameter estimation across two clock transitions is a route to mitigating systematic effects such as blackbody-radiation shifts, which currently limit clock accuracy.
• Tests of fundamental physics: Comparing two clock transitions is the standard approach to searching for variations of fundamental constants; joint estimation could improve the precision of such comparisons.
• Quantum-enhanced sensing networks: Entangled multi-level probes could improve distributed or multi-observable sensing where several quantities must be read out in one interrogation cycle.
• Qudit-based quantum information processing: The same magic trapping, qutrit control, qutrit entangling gate, and state-resolved readout are the ingredients for qudit quantum computing and novel quantum error-correcting codes.

Industry relevance. Neutral-atom arrays are a leading architecture for commercial quantum computing and for next-generation optical clocks. Demonstrating qutrit gates, qutrit entanglement, and qutrit readout in a clock-compatible platform is directly relevant to companies and institutes building either high-accuracy clocks or qudit-based processors, and the emphasis on noise-robust circuit design against realistic gate fidelities speaks to practical deployment.

Future Directions

• Extending to larger qutrit-entangled states. The authors suggest combining large atomic arrays with coherent atom shuttling or analog evolution to generate multi-partite qutrit entanglement, and note that lower-depth or noise-model-optimized circuits would widen the range of realistic advantage.
• Reaching useful dark times. The current demonstration uses zero dark time and does not characterize systematic effects. Coherence is limited by polarization gradients across the tweezer array that induce a differential light shift between ³P₂ and the other states — especially at the 500 µK trap depth used for Lamb-Dicke operation of the R₀₁ rotations. Recoil-free three-photon configurations, more homogeneous trap polarization, and reduced trap depth during the dark time are proposed remedies.
• Improving single-qutrit gate fidelity. Since the shortfall from the ideal Y_joint ≈ 0.5 traces to incoherent damping in the R₀₁ and R₁₂ rotations, feed-forward techniques to mitigate phase noise should directly improve joint estimation.
• Non-commuting parameters and more of them. Going beyond two commuting phases — to higher-dimensional states with more parameters, or to simultaneous estimation of phases and noise — is flagged as a harder and interesting open problem where saturating the quantum Cramér-Rao bound is more challenging.

Target Audience

Advanced readers in quantum metrology and neutral-atom quantum information: researchers building optical clocks or atomic-array processors, theorists working on multi-parameter estimation and qudit entanglement certification, and graduate students already comfortable with the quantum Cramér-Rao bound, Schmidt-number entanglement witnesses, and Rydberg-gate experiments. Experimentalists evaluating whether qutrit encoding is worth adding to their apparatus will find the gate characterization and noise-robust circuit design particularly useful. The paper is not suited to beginners, as it assumes fluency in both estimation theory and neutral-atom experimental technique.

Authors’ abstract

Quantum metrology harnesses entanglement to improve measurement precision beyond classical limits. While standard protocols rely on two-level qubits to estimate a single parameter, extending them to entangled multi-level qudits enables the optimal simultaneous estimation of multiple parameters within a single probe. However, generating such multi-level entanglement and harnessing it for joint multi-parameter estimation in atomic clocks has remained an outstanding challenge. Here, we experimentally demonstrate genuine qutrit entanglement and joint multi-parameter estimation in an optical clock platform. Leveraging control over the ground state and two fine-structure clock states of $^{88}\text{Sr}$ atoms trapped in triple-magic optical tweezers, we generate a maximally entangled two-qutrit state with a loss-postselected fidelity of F = 0.85(1), certifying genuine multi-level entanglement. Taking advantage of this high-dimensional entanglement, we theoretically construct and experimentally realize an optimal two-qutrit metrological probe state and noise-robust readout circuit to simultaneously estimate injected phases on two optical clock transitions. We observe a joint estimation variance below the ideal individual two-level sensing threshold, and show theoretically that this advantage persists at state-of-the-art atom numbers under circuit-level noise. These results demonstrate the key building blocks towards quantum information science with high-dimensional states encoded in the internal energy levels of neutral atoms.