Spontaneous Raman scattering out of a metastable atomic qubit

I. D. Moore, A. Quinn, J. O’Reilly, J. Metzner, S. Brudney, G. J. Gregory, D. J. Wineland, and D. T. C. Allcock · 2025

Read Spontaneous Raman scattering out of a metastable atomic qubit by I. D. Moore, A. Quinn, J. O’Reilly, J. Metzner, S. Brudney, G. J. Gregory, D. J. Wineland, and D. T. C. Allcock (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2505-04854.txt.

Introduction

Isaac Moore and co-authors measure a quiet error source for a promising kind of quantum bit. Trapped-ion computers store quantum bits in long-lived atomic states, which simplifies hardware and lets errors be flagged and erased. But the lasers that steer these atoms can scatter light off them, and that scattering sets the floor for how accurate control can be. The team measured the harmful Raman part of that scattering for a single trapped calcium ion.

They shone 976-nanometer light, far detuned from the atom's strong transition, on the ion's metastable qubit state and counted how often scattering knocked it out of that state. The rates they found feed directly into error budgets for one- and two-qubit operations. Their conclusion is encouraging: gate errors below one part in ten thousand look reachable on this platform, which supports metastable qubits as a serious route to large machines.

Catalog record

CREATORI. D. Moore, A. Quinn, J. O’Reilly, J. Metzner, S. Brudney, G. J. Gregory, D. J. Wineland, and D. T. C. Allcock
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE16 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2505-04854

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Spontaneous Raman scattering out of a metastable atomic qubit I. D. Moore, A. Quinn, J. O’Reilly, J. Metzner, S. Brudney, G. J. Gregory, D. J. Wineland, and D. T. C. Allcock Abstract Metastable qubits in atomic systems can enable large-scale quantum computing by simplifying hardware requirements and adding efficient erasure conversion to the pre-existing toolbox of high-fidelity laser-based control. For trapped atomic ions, the fundamental error floor of this control is given by spontaneous Raman and Rayleigh scattering from short-lived excited states. We measure spontaneous Raman scattering rates out of a metastable D 5 / 2 D_{5/2} qubit manifold of a single trapped 40 Ca + ion illuminated by 976 nm light that is -44 THz detuned from the dipole-allowed transition to the P 3 / 2 P_{3/2} manifold. This supports the calculation of error rates from both types of scattering during one- and two-qubit gates on this platform, thus demonstrating that infidelities < 10 − 4 <10^{-4} are possible. Stimulated Raman transitions are commonly driven to manipulate qubits encoded in trapped ions, neutral atoms [ 1 , 2 ] , defects in solid state systems [ 3 , 4 ] , and more [ 5 ] . Raman-based quantum logic gates on trapped-ion qubits have been demonstrated with fidelities \gtrsim 99.9\% [ 6 , 7 ] , sub- \mu s time scales [ 8 ] , absolute phase stability across space and time [ 9 ] , straightforward individual qubit optical addressing [ 10 ] , and 3- and 4-body interactions [ 11 ] . The fundamental error floor for these gates comes from spontaneous Raman and Rayleigh scattering through the short-lived P_{1/2} and P_{3/2} manifolds, which is already the limiting error source in the highest-fidelity experiments [ 6 , 7 ] . This limit can be reduced by increasing the Raman laser detuning while maintaining similar Rabi frequencies with a corresponding increase in laser power [ 12 ] . Together with continued improvements to qubit and motional coherence, this will enable higher-fidelity operation of trapped-ion quantum computers and reduce the required overhead for fault-tolerant quantum computation. Recent work [ 12 ] expanded a model for spontaneous scattering rates of ground-state qubits in trapped ions [ 13 ] to also include qubits encoded in the metastable D_{5/2} manifolds. Further, this theory extended beyond the small detuning regime by considering the frequency dependence of the density of states. The updated model has been experimentally verified for ground-state qubits in 133 Ba + [ 14 ] and predicts no fundamental error floor for ground-state qubits as the detuning between the Raman laser wavelength and the atomic transition grows, contrary to the model presented in Ref. [ 13 ] . Encoding quantum information in metastable manifolds that are spectrally isolated from fluorescence operations enables mid-circuit recooling and measurement [ 15 , 16 , 17 ] and photonic interconnects for modularity [ 18 ] without cumbersome dual-species operation [ 19 , 20 , 21 ] . Furthermore, metastable qubits have a number of unique advantages, including heralded state preparation [ 22 ] and the efficient conversion of leakage errors into less harmful erasure errors [ 23 , 24 , 25 ] . Recent demonstrations of entanglement between qubits encoded in the metastable D_{5/2} manifolds of trapped barium [ 26 , 27 ] and calcium [ 25 ] ions have utilized laser-driven stimulated Raman transitions. Our work on high-fidelity logic gates in calcium [ 25 ] took advantage of well-established laser technology at 976 nm to achieve high optical power in a compact package and at a wavelength that is -44 THz detuned from resonance with the dipole-allowed D_{5/2}\leftrightarrow P_{3/2} transition at 854 nm. In this work, we measure the spontaneous scattering rate out of the metastable D_{5/2} manifold of a 40 Ca + ion induced by 976 nm laser radiation. We confirm the theoretical prediction of Ref. [ 12 ] , thus demonstrating that spontaneous Raman scattering errors can be suppressed to <10^{-4} per two-qubit entangling gate at this wavelength. These conclusions are further supported by parallel work [ 28 ] that measured scattering rates out of metastable barium ion qubits at multiple wavelengths. I Experimental setup I.1 Detection Scheme We trap a single 40 Ca + ion in a four rod rf Paul trap [ 29 ] and Doppler cool it by applying a 397 nm beam with a projection onto all three principal axes of the trap and an 866 nm beam blue-detuned from the D_{3/2}\leftrightarrow P_{1/2} transition to repump out of the metastable D_{3/2} manifold while avoiding dark resonances [ 30 ] , see Figure 1 (a). We encode our qubit in the \lvert\uparrow\rangle\equiv\lvert D_{5/2},m_{J}=+5/2\rangle and \lvert\downarrow\rangle\equiv\lvert D_{5/2},m_{J}=+3/2\rangle metastable states and prepare the \lvert\uparrow\rangle state via polarization-selective optical pumping [ 25 ] . A quantization field of 1.56 G separates adjacent sublevels of the D_{5/2} manifold by 2.63 MHz via the linear Zeeman effect. All dipole-allowed transitions out of this manifold are detuned by at least several THz from the 397 nm and 866 nm beams; therefore an ion “shelved” in D_{5/2} will not scatter any light from the Doppler cooling beams [ 31 ] . We can thus measure leakage out of D_{5/2} by applying these beams and detecting 397 nm fluorescence on a photo-multiplier tube, a technique we will refer to as a “fluorescence check.” Figure 1: Energy level diagram for 40 Ca + . The fluorescence cycle (blue) is spectrally isolated from the qubit manifold, enabling us to detect population outside the D_{5/2} manifold using standard fluorescence collection techniques. We use this to measure decay (green) in the D_{5/2} population due to its finite lifetime and scattering of the 976 nm beam (red). Finally, we can depopulate D_{5/2} by applying a resonant 854 nm beam (orange). To prepare the \lvert\downarrow\rangle state after optically pumping to \lvert\uparrow\rangle , we apply a near-resonant (-21.6 GHz) \sigma^{+} -polarized 854 nm beam that breaks the D_{5/2} manifold splitting degeneracy [ 32 ] , see Fig. 2 . Then, we use a resonant rf drive applied to one of the trap rods to generate an oscillating magnetic field and transfer population to \lvert\downarrow\rangle . At the end of an experiment, we can distinguish between the qubit states by similarly driving population from \lvert\downarrow\rangle to \lvert s\rangle\equiv\lvert D_{5/2},m_{J}=+1/2\rangle and applying a resonant \pi -polarized 854 nm beam that pumps the D_{5/2} sublevels with |m|<5/2 into the S_{1/2} and D_{3/2} manifolds with >99\% fidelity. If the ion is dark during the subsequent fluorescence check, we apply another resonant 854 nm beam with both \sigma^{+} and \sigma^{-} components to completely depopulate the D_{5/2} manifold. If the ion is still dark after this, we assume that it either decrystallized 1 1 1 Significant heating events, most often collisions with background gas molecules, can disrupt the ion’s coupling to laser beams and reduce detectable fluorescence to background levels. or left the trap and we discard the measurement result. Figure 2: Qubit definition and differential light shifting. The 2.63 MHz separation between each D_{5/2} sublevel is suppressed for clarity and other energies are not to scale. The \sigma^{+} -polarized 854 nm beam -21.6 GHz detuned from resonance with the D_{5/2}\leftrightarrow P_{3/2} transition induces an AC Stark shift on

Read the complete text