{"id":"24bbd592-89c8-49a0-9a51-dded5d749e7c","arxiv_id":"2505.22466","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"Measurements of spontaneous Raman scattering from metastable 137Ba+ states agree with the Moore et al. model, supporting predictions of low gate errors at large detunings.","lead":"This paper measures how often lasers accidentally kick a metastable-state barium ion out of its qubit state, testing a recent theory of Raman scattering errors. If the theory holds, two-qubit gate error rates near 10^-4 may be reachable with far-detuned light, which matters for trapped-ion quantum computing.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Gate-error claim extrapolates from measured S1/2/D3/2 leakage to unmeasured D5/2-to-D5/2 scattering; the data do not test the channel that sets the far-detuning lower bound.","rationale":"The reader's weakest_assumption emphasized the AC Stark shift intensity calibration and the 461-nm polarization assumption. Those are legitimate experimental systematics, but the slope comparison in Fig. 3 is formed from the ratio of scattering rate to differential Stark shift, so common matrix-element errors partially cancel, and the quoted polarization uncertainty is small. The more load-bearing issue for the central claim is the mismatch between what is measured (out-of-manifold leakage) and what the 1e-4 gate-error number requires (total scattering, including within-D5/2 channels). The paper is explicit that back-scattering into D5/2 is not measured and that Table II includes it via theory, so the headline error estimate is an extrapolation. This does not overturn the paper's main experimental result, which is a clean measurement of the leakage channels with good agreement to the Moore model, but it does mean the far-detuning gate-error claim is conditional on the untested within-manifold part of the model. A feasible follow-up measurement of population transfer within the D5/2 manifold would settle the point; if the within-manifold rate matches the Moore prediction, the 1e-4 claim is supported, and if it saturates at the Ozeri-style lower bound, the central claim would need revision. Because the reader already assigned a conditional verdict and the new concern sharpens rather than reverses that conditionality, the verdict should remain UNCHANGED.","tokens_in":9347,"tokens_out":20808,"duration_ms":261980,"concrete_test":"Using the same trapped-ion setup, prepare |0> = |5D5/2,F=1,mF=0>, apply 674-nm and 461-nm scattering light for a variable time, then shelve the remaining 5D5/2 population back to 6S1/2 with the 1762-nm quadrupole transition and measure fluorescence, using microwave or Raman pulses to resolve individual D5/2 hyperfine or Zeeman sublevels before shelving. Comparing the inferred within-D5/2 scattering rate to the Moore-model value used in Table II settles whether the unmeasured channel is negligible; if it is not, the 1e-4 gate-error claim lacks direct experimental support.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim is that the measurements validate the Moore et al. model and support approximately 1e-4 metastable gate errors at tens of THz detuning. What is actually measured is the loss rate out of |0> = |5D5/2,F=1,mF=0> into the 6S1/2 and 5D3/2 manifolds only, as stated in the Fig. 3 caption ('the sum over f includes all states in 6S1/2 and 5D3/2'). The SRS error budget for a Raman gate, however, includes scattering into all states, in particular other levels within the 5D5/2 manifold, i.e. Rayleigh-like elastic scattering and inelastic transfer between qubit sublevels. The paper itself notes in Sec. IV that 'scattering back into D5/2 (excluding |0>) would need to be included,' and Table II adds this contribution by scaling the measured values with the very model under test. The far-detuning controversy between Ozeri et al. [15] and Moore et al. [16] concerns whether the total scattering error keeps decreasing with detuning, and the within-D5/2 channels are precisely where an Ozeri-style lower bound would appear. Both model curves in Fig. 3 omit these channels, and eq. (A.7) likewise excludes them. Thus the agreement shown in Fig. 3 validates only the leakage terms, not the full scattering model underlying the 1e-4 gate-error projection. The paper is transparent about this limitation, but the abstract and conclusion present the 1e-4 number as supported by the measurement; that support is indirect and depends on untested parts of the model.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"This manuscript reports measurements of spontaneous Raman scattering out of the metastable |5D5/2,F=1,mF=0> state of 137Ba+ at 617 nm (near resonance), 674 nm (red-detuned), and 461 nm (blue-detuned). The scattering rate per unit differential AC Stark shift is extracted from the shortening of the D5/2 lifetime under illumination, and the results are compared with the Ozeri et al. and Moore et al. models. The authors report better agreement with Moore et al. and use this agreement to argue that metastable-state two-qubit Raman gates with errors near 10^-4 are achievable at tens of terahertz of detuning.","tokens_in":9688,"tokens_out":5986,"duration_ms":68933,"significance":"The experimental method is sound and the data set is valuable: three wavelengths, red and blue detuning, direct lifetime-based rate extraction, and a comparison with two published models using the differential AC Stark shift as an intensity-independent x-axis. If the claim is restricted to scattering into 6S1/2 and 5D3/2, the data indeed favor the Moore model over Ozeri at large detuning, in line with the prior ground-state measurement in 133Ba+. However, the headline gate-error projection requires the unmeasured within-D5/2 scattering channels, so the significance is conditional. The manuscript is transparent about this limitation in Sec. IV, but the abstract and conclusion do not carry the same caveat.","major_comments":[{"comment":"The measurement and the model comparison in Fig. 3 are restricted to final states in 6S1/2 and 5D3/2, as the caption explicitly states. This validates only the leakage contribution to SRS, not the total scattering error, which also includes Rayleigh and Raman processes ending in the 5D5/2 manifold. The Ozeri-Moore disagreement at large detuning concerns the total scattering error, and the within-D5/2 channels are precisely where an Ozeri-style lower bound would appear. The paper acknowledges this in Sec. IV, noting that 'scattering back into D5/2 (excluding |0>) would need to be included,' and Fig. 4 and Table II include this channel only through model-dependent scaling. Therefore the abstract's statement that the measurements suggest metastable-level two-qubit gates with error rate approximately 10^-4 are possible is not directly supported by the data. The authors should either measure or bound repopulation of the D5/2 manifold, or revise the central claim so that it explicitly applies only to leakage out of the qubit manifold.","section":"Fig. 3 caption; Sec. IV"},{"comment":"The 'Scaled experiment' column in Table II is not an experimental result: it is obtained by multiplying the measured leakage rate by factors that account for the |1> state population and for scattering back into D5/2, using the very model under test. No formula or uncertainty for this scaling is given. As a result, the two-qubit error numbers in Table II cannot serve as evidence discriminating between the Ozeri and Moore models. The conclusions should be rephrased so that the gate-error projection is explicitly identified as a model-based extrapolation rather than a directly measured quantity.","section":"Table II; Sec. IV"}],"minor_comments":[{"comment":"The 461-nm polarization is assumed equal to the 617-nm polarization based on a polarizing beam splitter measurement, but no uncertainty is quoted for the polarization difference; since the scattering rate depends on polarization, a quantitative bound on any residual difference would strengthen the 461-nm point.","section":"Sec. III"},{"comment":"The slopes of the linear fits in Fig. 3 are not reported numerically; providing them, with uncertainties, would make the model comparison quantitative and reproducible.","section":"Fig. 3"},{"comment":"There is a typo in 'scattering back into D5/2 (excluding |0>) would need to be be included,' and Ref. [19] contains 'and and NIST'; these should be corrected.","section":"Sec. IV"},{"comment":"The concluding claim that the SRS-induced error can be reduced below 10^-5 per single-qubit gate should carry the same caveat as the two-qubit projection: it relies on the model-dependent D5/2 back-scattering correction and is not a directly measured quantity.","section":"Conclusion"}],"recommendation":"major_revision","confidential_remarks":"The stress-test concern is real and is confirmed by the manuscript's own text: the measured channel set excludes D5/2 final states, and Table II repairs this with the model under test. The experimental core appears sound, so I do not recommend rejection; a revision that either measures/bounds the D5/2 repopulation or clearly reframes the abstract and conclusions as leakage-only would make the claims match the evidence."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: this is the first experimental SRS measurement on a metastable-state qubit, and the core data are solid. Measured leakage rates out of |0> = |5D5/2,F=1,mF=0> into 6S1/2 and 5D3/2 at 674 and 461 nm sit on the Moore et al. curve and clearly below the Ozeri-style prediction. The lifetime-vs-Stark-shift extraction is careful, the polarization is characterized for 674 and 617 nm, and the paper is honest that ladder terms are not tested. That part deserves a serious referee.\n\nThe soft spot is the gate-error extrapolation. What is measured is only scattering out of the D5/2 manifold; the experiment does not detect scattering between D5/2 sublevels, which is precisely the channel where an Ozeri-like lower bound would show up in the far-detuned limit. Table II folds that missing channel in by scaling the measured rates with the very model under test. The paper flags this in the Fig. 4 caption and the Table II note, so it is transparent, but the abstract and conclusion phrase the 10^-4 result as if the measurement supports it directly. The support is indirect and depends on untested parts of Moore's theory.\n\nMinor caveats: the 461 nm polarization is assumed equal to 617 nm, with no quoted uncertainty on that equality, and the Ozeri-extension in the appendix is a simplified version—fine for comparison, but don't mistake it for the full model. None of this undermines the measured leakage rates.\n\nWho should read it: trapped-ion experimentalists working on metastable or dual-type qubits, especially OMG-protocol people, and theorists interested in spontaneous scattering limits. It deserves peer review. My recommendation: accept with revisions that tighten the abstract/conclusion to distinguish measured leakage from modeled total scattering, and ideally add one sentence describing what is direct evidence versus model extrapolation.","headline":"Clean SRS leakage measurements in a metastable Ba+ qubit that favor Moore over Ozeri, but the headline 10^-4 gate-error numbers lean on unmeasured within-manifold scattering.","tokens_in":10239,"tokens_out":2773,"would_cite":true,"duration_ms":30802,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper reports that spontaneous Raman scattering from a metastable state of 137Ba+ follows the four-process model [16], so Raman gates detuned by tens of terahertz can reach error rates near 10^-4 with no fundamental scattering floor.","keywords":["spontaneous Raman scattering","metastable qubits","trapped-ion quantum computing","barium-137","stimulated Raman transitions","quantum gate error rates","AC Stark shift","atomic transition data"],"falsifier":"A direct check would calibrate the 674 nm intensity by an independent method, such as a power meter and measured beam waist, and compare the extracted scattering rate per unit intensity with the Stark-shift-based value; a common multiplicative shift would expose the Stark-shift assumption. The untested ladder terms could be probed by choosing initial and final states with $(E_i-E_f)>\\hbar\\omega_\\ell$ and a sufficiently red-detuned laser, where the two ladder channels are energy-conserving.","tokens_in":9155,"feed_emoji":"⚛️","tokens_out":7820,"duration_ms":73817,"temperature":0.7,"pith_summary":"Spontaneous Raman scattering is the fundamental error source for stimulated-Raman quantum gates in trapped ions, and the authors set out to test whether that error has a true lower bound. They measure the scattering rate out of a metastable $|{}^5D_{5/2},F=1,m_F=0\\rangle$ state of $^{137}$Ba$^+$ at 617 nm, 674 nm, and 461 nm, and compare the rates to two models. The data agree with the four-process model [16] and deviate from the earlier single-process model [15] at the large detunings where the predictions diverge. If correct, the result means far-detuned lasers can suppress Raman scattering errors below $10^{-4}$ per two-qubit gate, eliminating a supposed fundamental limit.","feed_headline":"Raman scattering data support no lower bound on metastable gate error","feed_subtitle":"Measured rates in 137Ba+ at 617, 674, and 461 nm match the four-process model, pointing to ~1e-4 two-qubit errors.","key_machinery":"The load-bearing object is the four-process scattering rate formula of [16], Eq. (1): a sum over intermediate $k$ and scattered-photon polarization $q$ of $\\Lambda$, $V$, and two ladder amplitudes, each weighted by its own detuning denominator and an energy-conserving Heaviside factor. The comparison harness is the differential AC Stark shift of Eq. (A.1), which converts measured laser intensity into a common scale and lets the experiment test the model without a separate intensity calibration. The ladder terms were not tested here because the energy-conservation condition $(E_i-E_f)>\\hbar\\omega_\\ell$ was not met for the chosen states.","core_discovery":"The central discovery is that spontaneous scattering from a metastable qubit level in $^{137}$Ba$^+$ behaves as the four-process formula Eq. (1) of [16] predicts, rather than as the simpler model [15] predicts. The experiment prepares $|0\\rangle = |{}^5D_{5/2},F=1,m_F=0\\rangle$, applies scattering light at three wavelengths, and extracts the decay rate into $6S_{1/2}$ and $5D_{3/2}$ by comparing the measured lifetime with and without the scattering beam. Laser intensity is calibrated through the differential AC Stark shift, so the data are plotted as scattering rate per unit Stark shift and compared directly with the theoretical curves. At 674 nm and 461 nm the measured slopes track the four-process model and fall below the older prediction, supporting the conclusion that the scattering rate can be reduced without a finite floor by increasing red or blue detuning.","pith_inferences":["If the four-process model holds at even larger detunings, gate engineering can trade speed against scattering error continuously, and the practical limit for Raman gates becomes available laser power and gate-time decoherence rather than a scattering floor.","The ladder-scattering channels in Eq. (1) have not been experimentally isolated; a measurement from a higher-energy metastable sublevel with a red-detuned beam could validate or falsify that part of the model.","The same comparison of scattering rate to AC Stark shift could be applied to metastable qubits in other species, such as Ca+ or Sr+, to test whether the four-process model is universal across different intermediate-state structures.","The results strengthen the case for dual-type encodings that use metastable states for operations and ground states for storage, since metastable-manifold Raman gates can be made scattering-clean without disturbing ground-state coherence."],"forward_implications":["Far-detuned Raman operations inside the metastable $D_{5/2}$ manifold can reach single-qubit spontaneous-scattering errors below $10^{-5}$ per gate.","Two-qubit entangling gates driven tens of terahertz from resonance can approach error rates near $10^{-4}$, limited by other decoherence sources rather than by scattering.","The measurements discriminate between the two scattering models, favoring the four-process description at large detunings and finding no evidence for a finite detuning floor.","The ratio of scattering rate to differential AC Stark shift provides a calibration-free way to compare theory and experiment for other metastable species."],"supporting_citations":[{"why":"supplies the four-process scattering rate formula Eq. (1) that the measurements are designed to test","marker":"[16]"},{"why":"provides the older single-process scattering model that serves as the baseline for comparison","marker":"[15]"},{"why":"reports ground-state measurements in 133Ba+ that already favored the four-process model and motivate extending the test to metastable states","marker":"[17]"},{"why":"provides the reduced matrix elements and lifetimes used in both the scattering-rate and AC Stark-shift sums","marker":"[18]"},{"why":"provides the atomic energy levels used in the detunings and scattered-photon frequencies","marker":"[19]"},{"why":"defines the Mølmer-Sørensen gate whose power scaling is used to convert single-qubit scatter rates to two-qubit error estimates","marker":"[24]"},{"why":"reports a parallel measurement in metastable 40Ca+ that independently supports the same conclusion","marker":"[25]"}],"fun_headline_variants":["Metastable Ba+ Raman scattering backs no-lower-bound gate error","Ba+ experiment validates no floor on Raman gate error","Far-detuned Raman in Ba+ may slash metastable gate errors","Metastable qubit test: no limit to error reduction from detuning","Scattering data in Ba+ support infinite detuning for low error"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The measurement assumes the laser intensity is correctly inferred from the differential AC Stark shift, which is computed by summing dipole matrix elements from the atomic-data reference [18]; a systematic error in those matrix elements or in the Stark-shift formula would shift every extracted scattering rate by the same factor.","fun_headline_variants_meta":{"raw":{"variants":["Metastable Ba+ Raman scattering backs no-lower-bound gate error","Ba+ experiment validates no floor on Raman gate error","Far-detuned Raman in Ba+ may slash metastable gate errors","Metastable qubit test: no limit to error reduction from detuning","Scattering data in Ba+ support infinite detuning for low error"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000285,"raw_usage":{"total_tokens":1646,"prompt_tokens":882,"completion_tokens":764,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":498,"completion_tokens_details":{"reasoning_tokens":673}},"tokens_in":498,"tokens_out":764,"duration_ms":7521,"temperature":1.0,"reasoning_tokens":673,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T13:04:50.303333+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A direct check would calibrate the 674 nm intensity by an independent method, such as a power meter and measured beam waist, and compare the extracted scattering rate per unit intensity with the Stark-shift-based value; a common multiplicative shift would expose the Stark-shift assumption. The untested ladder terms could be probed by choosing initial and final states with $(E_i-E_f)>\\hbar\\omega_\\ell$ and a sufficiently red-detuned laser, where the two ladder channels are energy-conserving.","supporting_citations":[{"cited_title":"B ˘az˘avan, S","cited_arxiv_id":null,"evidence_quote":"supplies the four-process scattering rate formula Eq. (1) that the measurements are designed to test"},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"reports ground-state measurements in 133Ba+ that already favored the four-process model and motivate extending the test to metastable states"},{"cited_title":"Ozeri, C","cited_arxiv_id":null,"evidence_quote":"provides the reduced matrix elements and lifetimes used in both the scattering-rate and AC Stark-shift sums"},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"provides the atomic energy levels used in the detunings and scattered-photon frequencies"},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"defines the Mølmer-Sørensen gate whose power scaling is used to convert single-qubit scatter rates to two-qubit error estimates"},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"reports a parallel measurement in metastable 40Ca+ that independently supports the same conclusion"}],"review_version":1}