Influence of fast and slow laser phase noise on the fidelity of the Mølmer-Sørensen trapped-ion gate

AI summary

Overview

Research area: trapped-ion quantum computing, specifically the Mølmer-Sørensen (MS) two-qubit entangling gate and its sensitivity to laser phase noise. Technical level: Advanced. Scope: a theoretical derivation of the noise Hamiltonians, filter functions, and average gate-error expressions for fast and slow laser phase noise in the two common MS laser-beam geometries.

What This Paper Is About

High-fidelity two-qubit entangling gates are needed for useful trapped-ion quantum processors, and the MS gate is a leading choice because of its resilience to ion temperature and demonstrated record fidelities. The problem is that the driving laser is not spectrally pure: phase noise creates parasitic transitions that reduce gate fidelity. This paper identifies the two noise frequency ranges that matter most and derives analytical expressions for the resulting average gate error in terms of the laser noise power spectral density and the gate parameters.

Key Contributions

1. It identifies two dominant noise bands for the MS gate: “fast” noise near the motional mode spectrum, and “slow” noise at frequencies on the order of the inverse gate time.
2. It derives the noise Hamiltonians for both common MS beam geometries, the phase-sensitive configuration with co-propagating bichromatic components and the phase-insensitive configuration with counter-propagating components.
3. It obtains analytical filter functions for the fast and slow noise ranges, showing that the average gate error is an overlap integral of the laser phase or frequency noise power spectral density with a filter function that depends only on gate parameters.
4. It provides simplified error estimates for slowly varying noise spectra and validates the findings against previously published numerical simulations.

Main Findings

• Two spectral ranges dominate. Fast noise lies near the motional mode spectrum, from hundreds of kHz to several MHz, with timescales around 1 μs and less. Slow noise lies at frequencies on the order of μ − ω_l, in the kHz to tens of kHz range, with a timescale of roughly 0.1–1 ms.
• The physical sources differ by band. Fast noise is linked to the servo loop used to lock the laser, including servo bumps near the servo bandwidth in the MHz range, and to relaxation oscillations of the gain medium, whose peak lies in the sub-MHz range for solid-state lasers. Slow noise is linked to acoustic and thermal effects, mechanical properties of the reference cavity, 1/f frequency noise, and flicker noise, with a cutoff frequency of 10–100 Hz before white noise dominates.
• Common-source phase noise enters as a spin-phase fluctuation. Because all bichromatic components usually come from one laser source, their phase fluctuations are identical. The motional phase is unaffected, while the spin phase acquires the source phase noise, so δφ_S(t) = δφ(t) and δφ_M(t) = 0.
• Noise Hamiltonians are derived for both geometries. The phase-sensitive configuration uses the “++” sign choice and the phase-insensitive configuration uses one “+” and one “−”. In each case the fast noise term is carrier-like and the slow noise term is sideband-like; the two ranges are treated independently because their spectral overlap is insignificant.
• The average gate error is a filter-function integral. For fast noise, the error is written as an integral of the one-sided phase noise power spectral density multiplied by a filter function. For slow noise, the corresponding frequency noise power spectral density enters. The filter function depends on gate parameters rather than on the noise spectrum.
• Fast-noise filter function is peaked at the detuning. For constant laser power, equal Rabi frequency on both ions, and μτ ≫ 1, the phase-sensitive fast filter function is ℱ_fast^(S)(f) = [2(Ωτ)^2/5] sinc^2[((2πf − μ)τ)/2]. It is centered at μ/2π. For shaped pulses, the filter function is proportional to the squared τ-windowed Fourier transform of the pulse function and is smeared by approximately 1/τ.
• Slow-noise estimates are provided. The paper derives slow-noise filter functions and gives simplified error estimates for slowly varying noise spectra, describing them as a rule of thumb for setting laser spectral-purity requirements from a target gate error. The excerpt does not include the explicit final slow-noise formula.
• State averaging includes thermal motion. The mean error is averaged over random two-qubit starting states and a thermal motional state. The derivation uses the diagonal matrix element of the displacement operator in terms of Laguerre polynomials and a generating function to sum over motional occupations.
• No measured fidelities or experimental datasets are reported. The paper is theoretical and validates against previously published numerical simulations, not new experimental data.

Methodology in Plain English

The authors start from the standard MS gate Hamiltonian for two laser-beam geometries and assume operation in the Lamb-Dicke regime, equal intensities of the bichromatic components at the ion, and a single laser source supplying all fields. They express the laser phase fluctuations as small changes in the spin operators, which produces a noise Hamiltonian for each geometry. The noise is split into a fast carrier-like part and a slow sideband-like part. Using first-order perturbation theory in the form of a Dyson-series correction, they compute the average gate error over random qubit states and a thermal motional distribution. The noise correlation is converted into a power spectral density through the Wiener-Khinchin theorem, which yields the filter-function form of the error. Analytical integrals and known polynomial identities give closed expressions for the filter functions and simplified estimates in limiting cases. The results are then compared with previously published numerical simulations.

Why This Matters

This work gives trapped-ion experimentalists a direct way to connect laser noise spectra to MS gate error without running full numerical simulations for every noise spectrum. It clarifies which frequency bands deserve the most attention and provides analytical filter functions that can be used to set laser stabilization requirements, interpret measured noise spectra, and compare beam geometries.

Real-world applications:

• Setting phase-noise and frequency-noise specifications for lasers used in trapped-ion quantum processors.
• Designing laser stabilization servo loops and choosing laser sources while accounting for servo bumps and relaxation oscillations.
• Comparing co-propagating and counter-propagating MS gate geometries for sensitivity to laser noise.
• Building error budgets and noise diagnostics for scaling trapped-ion quantum computers.

Industry relevance: the results are relevant to quantum-computing hardware companies using trapped ions, laser and photonics manufacturers supplying stable sources, and metrology groups working on narrow-linewidth lasers and frequency stabilization.

Future Directions

• Experimental validation of the predicted filter functions and error estimates on trapped-ion MS gate setups.
• Extending the filter-function approach to shaped and modulated pulses, and using it to optimize pulse sequences against measured laser noise.
• Combining fast and slow laser phase noise with other error sources such as motional heating, spontaneous emission, and control errors.
• Applying the framework to multi-ion registers and comparing laser-noise requirements across different trapped-ion gate implementations, including Raman and laser-free schemes.

Target Audience

This paper is most useful to researchers and graduate students working on trapped-ion quantum computing, especially those modeling or measuring MS gate fidelity and laser noise. It also benefits laser stabilization engineers, quantum-hardware developers building error budgets, and theorists interested in filter-function methods for noisy quantum control.

Authors’ abstract

High-fidelity two-qubit entangling gates are essential for the realization of useful quantum algorithms on quantum processors. The Mølmer-Sørensen (MS) gate has become a common choice for trapped-ion quantum computing due to its resilience to ion temperature and its demonstrated record fidelities. However, the spectral impurity of the driving laser field impacts gate performance, with phase noise influencing the qubit dynamics through mechanisms operating on different timescales. In this work, we present a comprehensive theoretical analysis of laser phase noise in the MS gate, identifying two spectral ranges which influence the gate fidelity the most: "fast" noise at frequencies near the motional mode spectrum, and "slow" noise at frequencies on the order of the inverse gate time. We derive the noise Hamiltonians for two common laser beam geometries and obtain analytical expressions for the average gate error in terms of the laser noise power spectral density and gate parameters. For slowly varying noise spectra, we provide simplified error estimates. In addition, we validate our findings against previously published numerical simulations.