{"id":"fce891f2-ba02-4a90-9b3e-0140ce4d4205","arxiv_id":"2505.04854","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"Measured spontaneous Raman scattering rates out of the metastable D5/2 qubit manifold of a single 40Ca+ ion under 976 nm light, confirming the large-detuning scattering model and supporting sub-1e-4 gate error estimates.","lead":"Researchers measured how often 976 nm laser light knocks a trapped calcium ion out of its long-lived qubit states through spontaneous Raman scattering. The measured rates match theory and imply that Raman logic gates on this platform can operate below one error in ten thousand.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The π-polarization dataset may be biased by discarding |↑> trials, so the Table I agreement is less secure than stated.","rationale":"The reader's weakest assumption was that the decay analysis assumes no optical pumping among D5/2 sublevels and a constant single-exponential rate. My concern is a concrete failure mode of that assumption: the π dataset explicitly encounters scattering into |↑>, and the paper's stated mitigation (discarding those trials) does not yield a pure exponential with rate λ_SD, so the extracted Γ_SD for that configuration can be biased. This is a load-bearing issue for the full Table I validation, but it does not overturn the central conclusion because the two σ− measurements are unaffected and the two-qubit gate-error estimate is based on σ− scattering. The reader's conditional verdict, driven primarily by data availability, is appropriate; my concern strengthens the need for the raw data and for a corrected analysis of the π point. If the reanalysis confirms the bias is smaller than the quoted uncertainty, the agreement stands. If not, the π row of Table I would need revision, but the sub-1e-4 gate-error claim would likely remain intact. I therefore recommend no change to the reader's CONDITIONAL verdict, while flagging a specific systematic that should be checked before the three-point agreement is taken as full validation of the theory.","tokens_in":8637,"tokens_out":21071,"duration_ms":219061,"concrete_test":"Obtain the raw π-polarization dataset from the authors and reanalyze it without discarding |↑> trials: define Pop_D5/2(t) = (N_↓+N_↑+N_s)/N_total and fit the exact two-channel model dP_↓/dt = -(λ_SD+λ_up)P_↓, dP_↑/dt = λ_up P_↓, to extract λ_SD. Equivalently, simulate the censored analysis using the known branching ratio to |↑> and check whether the single-exponential fit of the censored survival fraction deviates from the true λ_SD by more than the reported 0.04 Hz/(nW/m2) uncertainty.","verdict_should_be":"UNCHANGED","load_bearing_attack":"For the π-polarized measurement from |↓>, population can scatter to |↑> (m=+5/2), a D5/2 state that no longer interacts with the π beam. The paper says it 'disregard[s] trials where we detect |↑> population' to mitigate this. This censoring changes the fitted quantity from the true D5/2 survival probability to P_↓/(P_↓+P_SD), where P_↓ is the population still in |↓> and P_SD is the population that left to S1/2/D3/2. If λ_SD is the desired loss rate to S/D and λ_up is the transfer rate to |↑>, this censored fraction equals e^{-(λ_SD+λ_up)t} / [e^{-(λ_SD+λ_up)t} + (λ_SD/(λ_SD+λ_up))(1-e^{-(λ_SD+λ_up)t})]. Its logarithmic derivative is λ_SD at t=0 but increases toward λ_SD+λ_up at later times. Fitting the single exponential of Eq. (1) over delays up to 1 s therefore overestimates λ_SD by an amount that depends on λ_up/λ_SD. The manuscript does not quantify λ_up or correct for this non-exponentiality. The branching from P3/2 to D5/2 is 5.88%, so λ_up/λ_SD is plausibly a few percent, comparable to the 2.8% statistical error quoted for the π point (1.43(4) vs theory 1.44(3)). Thus the agreement in Table I could be partly fortuitous for this configuration. The σ− numbers are not affected by this particular censoring issue, so the central gate-error estimate, which uses σ− quantities, would survive even if this bias is confirmed. The concern is that the claimed validation of the theory is weakened if one of the three headline points is systematically off.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper reports measurements of spontaneous Raman scattering rates out of the metastable D5/2 qubit manifold of a single trapped 40Ca+ ion illuminated by 976 nm light, for three combinations of initial Zeeman sublevel and beam polarization. The measured rates per unit intensity are compared to a parameter-free prediction from the authors' prior theory [12] and agree within uncertainties. The authors then combine these rates with known branching ratios to estimate spontaneous Raman and Rayleigh scattering errors for single- and two-qubit gates, concluding that infidelities below 1e-4 are achievable. The experimental work includes state preparation via optical pumping, fluorescence-based state detection, AC Stark shift intensity calibration, and subtraction of the natural D5/2 lifetime.","tokens_in":8980,"tokens_out":12658,"duration_ms":116766,"significance":"If the results hold, they provide the first experimental validation of the extended photon-scattering model of Ref. [12] for metastable qubits at large detuning, and they support the practical claim that sub-1e-4 error rates are possible for two-qubit gates on this platform. The measurements are careful: the natural lifetime is measured independently, the intensity calibration is tied to the AC Stark shift, and the statistical uncertainties are small (2.3-3%). The agreement of the two σ− data points with theory is very good. The main weakness is the treatment of the π-polarized dataset, where a dark-state censoring effect may bias the extracted rate, and the calibration procedure for the absolute intensity is not fully specified.","major_comments":[{"comment":"The procedure of disregarding trials where the ion is detected in |↑⟩ changes the fitted quantity from the total D5/2 survival probability to a censored fraction that is not a single exponential. The log-derivative of this fraction starts at Γ_SD and increases toward Γ_SD+Γ_up, where Γ_up is the transfer rate to |↑⟩. With the branching ratio to |↑⟩ of about 3.9% of P3/2 decays (Fig. 4(c)), the bias in the extracted Γ_SD is on the order of a few percent, comparable to the 2.8% statistical uncertainty of the (+3/2, π) point. The manuscript neither quantifies nor corrects for this effect. Please reanalyze the π dataset using the censored survival likelihood, or restrict the fit to times where the bias is negligible, and update Table I accordingly.","section":"Section II, π-polarized measurement and Table I"},{"comment":"The quoted gate-error estimates 1.25(2)×1e-6 and 5×1e-5 are calculations, not measurements, and the manuscript does not provide the explicit scaling from the measured Γ_SD to Γ_Ram, nor the assumed Rabi frequency, beam geometry, and Lamb-Dicke parameter. Without this information, the reader cannot reproduce the claimed error rates. Please include the full derivation and parameter values used.","section":"Section II A, Eqs. (3) and (4)"}],"minor_comments":[{"comment":"Please specify how the AC Stark shift measurement is converted to an absolute intensity, and whether the polarizability used is derived from the same theory under test; this would clarify the independence of the calibration from the theory being validated.","section":"Section I C, intensity calibration"},{"comment":"The arXiv number for Ref. [25] appears to contain a typo ('241.12727'); it should likely be '2411.12727'.","section":"References, Ref. [25]"},{"comment":"The double-scattering corrections of ≈0.6% and ≈1.4% are introduced without derivation; please provide the basis for these estimates or a reference.","section":"Section II, double-scattering corrections"},{"comment":"The branching percentages in Fig. 4(a-c) are not explicitly connected to the measured Γ_SD values in the text; a sentence explaining how these percentages relate to the fitted rates would improve clarity.","section":"Fig. 4 caption and Section II"},{"comment":"The abstract states that the measurements support calculations of error rates from 'both types of scattering', but the Rayleigh scattering bounds are taken from prior work (Refs. [36] and [12]); please clarify the provenance of those bounds.","section":"Abstract and Section II A"}],"recommendation":"major_revision","confidential_remarks":null},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nHere's the short version: this paper reports the first direct measurement of spontaneous Raman scattering rates out of the D5/2 metastable qubit manifold of 40Ca+ at 976 nm, and the three measured rates agree with the parameter-free prediction of Ref. [12] within uncertainties. It's a clean validation. The sub-1e-4 two-qubit gate error is an extrapolation using the same theory, not a direct gate measurement, but the underlying data are solid.\n\nThe experimental work is careful. The shelving + fluorescence check is standard, they subtract the natural lifetime measured in the same setup, calibrate beam intensities via AC Stark shifts, and the error bars include both statistical and theoretical (P3/2 lifetime) contributions. The three rates in Table I are 3.65(11), 2.13(5), 1.43(4) vs theoretical 3.60(6), 2.16(4), 1.44(3). That's a strong confirmation, and it extends the existing model [12] to the metastable case.\n\nSoft spots are minor but worth noting. First, no raw data are deposited; \"available upon reasonable request\" is a reproducibility red flag in 2025. Second, the π-polarization data point is obtained by discarding trials where the ion scattered into |↑>, the m=+5/2 state. As the stress-test worry notes, that changes the fitted quantity from the true D5/2 survival probability to a conditional probability, which biases the extracted rate upward. However, the paper's own branching numbers say only ~0.4% of scattering events from |↓> under π light go to |↑>, so the bias is likely around 0.4%, below the 2.8% statistical error on that point. The paper should have quantified the correction, but it doesn't threaten the result. The central gate-error estimate uses the σ− configuration, which doesn't have this issue.\n\nThe citation pattern is heavy on the authors' own papers, but the prediction being tested is parameter-free and made before this experiment, so the self-citation is not a flaw. The parallel barium work [28] is listed as in preparation, which is fine but not checkable.\n\nWho should read this: trapped-ion experimentalists working with metastable qubits, and anyone estimating Raman-scattering error floors. It deserves a serious referee. I'd send it to review, with a request for raw data or at least a more detailed description of the π-point analysis. If the data appear, the conditional status should lift to accept.","headline":"Careful measurement of Raman scattering from metastable Ca+ qubits confirms parameter-free theory; the π-point has a small unquantified bias but the central result holds.","tokens_in":9552,"tokens_out":8472,"would_cite":true,"duration_ms":79524,"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":"Spontaneous Raman scattering out of a metastable calcium qubit matches theory, keeping two-qubit gate errors below $10^{-4}$.","keywords":["spontaneous Raman scattering","metastable qubits","trapped ions","calcium-40","Raman gate errors","D5/2 manifold","erasure conversion"],"falsifier":"Shine $\\sigma^+$-polarized 976 nm light on the $|\\uparrow\\rangle$ state: the model says no dipole-allowed transition to $P_{3/2}$, so any measured scattering rate above noise would falsify the claim. A second check is to measure the $D_{5/2}$ decay curve at several intensities well beyond 1 s or with sublevel-resolved readout; a visible deviation from single-exponential decay or a rate that changes with intensity would indicate optical pumping or multiple-scattering effects the analysis neglects.","tokens_in":8435,"feed_emoji":"⚛️","tokens_out":7939,"duration_ms":68567,"temperature":0.7,"pith_summary":"This paper measures how often 976 nm Raman laser light knocks a qubit stored in the metastable $D_{5/2}$ manifold of a single trapped $^{40}$Ca$^+$ ion out of that manifold through spontaneous Raman scattering. The measured per-intensity rates agree with the parameter-free predictions of a recent scattering model to within the error bars, across three combinations of initial state and beam polarization. If the agreement holds, spontaneous Raman scattering is not a serious barrier for this qubit platform: the predicted error per two-qubit entangling gate is about $5\\times 10^{-5}$, below the $10^{-4}$ level often quoted as a practical floor. That matters because metastable-qubit architectures promise erasure conversion and mid-circuit operations, and the scattering rate is the fundamental error limit of their laser-based gates.","feed_headline":"Ion qubit scattering low enough for sub-1e-4 gates","feed_subtitle":"Measured rates match parameter-free theory, clearing the way for metastable-calcium Raman gates.","key_machinery":"The load-bearing quantity is the spontaneous Raman scattering rate $\\Gamma_{SD}$ per unit intensity out of a chosen $D_{5/2}$ sublevel, extracted from a single-exponential fit to shelved-population decay and the subtraction $\\Gamma_{SD} = 1/\\tau_{\\mathrm{meas}} - 1/\\tau_{\\mathrm{nat}}$. The measurement uses a fluorescence check: an ion outside $D_{5/2}$ scatters 397 nm cooling light, so decay of the dark fraction directly reports scattering out of the manifold. The comparison target is the model of Ref. [12], which computes scattering amplitudes with Clebsch-Gordan coefficients and a frequency-dependent density of states for the final photon, and the error conversion uses the gate-time formulas $P_{1q} = (\\pi\\Gamma_{\\mathrm{Ram}})/(2|\\Omega_R|)$ and $P_{2q} = (2/\\eta)\\cdot(\\pi\\Gamma_{\\mathrm{Ram}})/(2|\\Omega_R|)$.","core_discovery":"The paper claims that spontaneous Raman scattering rates out of the $D_{5/2}$ qubit manifold of $^{40}$Ca$^+$ under far-detuned 976 nm light agree with the density-of-states-corrected model of Ref. [12], at least for the $\\sigma^-$- and $\\pi$-polarized beams tested. The extracted per-intensity scattering rates to $S_{1/2}$ and $D_{3/2}$ are $3.65(11)$, $2.13(5)$, and $1.43(4)$ Hz/(nW/m$^2$) for the $(|\\uparrow\\rangle, \\sigma^-)$, $(|\\downarrow\\rangle, \\sigma^-)$, and $(|\\downarrow\\rangle, \\pi)$ configurations, against theoretical values $3.60(6)$, $2.16(4)$, and $1.44(3)$. From these rates and known branching ratios, the paper derives a single-qubit scattering error of $1.25(2)\\times 10^{-6}$ per gate and a two-qubit $\\sigma_z\\otimes\\sigma_z$ scattering error of $5\\times 10^{-5}$ at a 2 MHz secular frequency, with Rayleigh-induced errors below $10^{-7}$ or $10^{-8}$. The claim is that the error floor from spontaneous scattering in this platform is below $10^{-4}$ per entangling gate, contradicting the older model that predicted a larger floor at these detunings.","pith_inferences":["If the detuning trend continues, shifting the Raman wavelength beyond 976 nm should push the scattering floor below $10^{-5}$, a prediction that can be tested with the same apparatus.","The single-exponential decay assumption could be stress-tested by preparing specific $D_{5/2}$ sublevels and monitoring their population fractions during illumination; a sublevel-dependent rate would show up as a non-exponential total decay.","Measuring the same rates in a second metastable species, as the paper notes is underway, would test the generality of the theory across alkali-earth-like ions.","With scattering errors reduced, the next hurdles for these gates will be motional decoherence and laser-intensity noise, so combining this result with improved motional control is the natural route to the predicted fidelity."],"forward_implications":["At 976 nm, spontaneous Raman scattering contributes roughly $5\\times 10^{-5}$ error per two-qubit gate on metastable-calcium qubits, leaving room below the $10^{-4}$ fault-tolerance-relevant floor.","Single-qubit scattering errors are about $1.25(2)\\times 10^{-6}$ per gate, so the scattering floor is not the dominant limit for one-qubit operations at this wavelength.","Rayleigh scattering contributes below $10^{-7}$ per single-qubit gate and below $10^{-8}$ per two-qubit gate, effectively negligible against Raman scattering.","The validated theory says that increasing the Raman detuning further reduces the scattering error without a fundamental floor, in contrast to the older scattering model.","Because measured rates match theory without free parameters, the same fluorescence-check method can be used to characterize scattering for other metastable ion species and wavelengths."],"supporting_citations":[{"why":"Supplies the parameter-free scattering-rate model for metastable $D_{5/2}$ qubits that the measured rates are compared against.","marker":"[12]"},{"why":"Provides the spontaneous-scattering gate-error formula used to convert measured rates into single-qubit error probabilities.","marker":"[13]"},{"why":"Supplies the natural $D_{5/2}$ lifetime used to extract the induced scattering rate from the measured decay constant.","marker":"[33]"},{"why":"Supplies the $P_{3/2}$ lifetime whose uncertainty dominates the theoretical error bars on the predicted rates.","marker":"[35]"},{"why":"Establishes the 976 nm Raman gate platform in calcium whose error floor is being characterized.","marker":"[25]"},{"why":"Prior verification of the theory for ground-state qubits in barium, motivating the metastable-manifold test performed here.","marker":"[14]"},{"why":"Supplies the $P_{3/2}$ branching fractions used to convert measured rates into scattering rates back to the $D_{5/2}$ manifold.","marker":"[37]"}],"fun_headline_variants":["Raman scattering rates confirm sub-1e-4 qubit gate errors","Metastable qubit scattering rates clear path to 1e-4 gates","Scattering out of Ca+ qubit measured, gate infidelity below 1e-4","Calcium ion qubit scattering rates enable ultra-low error gates","Ion qubit scattering measured: sub-1e-4 gate errors feasible"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The data analysis assumes the $D_{5/2}$ population decays as a single exponential with a constant rate, that the 976 nm beam does not pump population among $D_{5/2}$ sublevels, and that double-scattering corrections of 0.6% to 1.4% are negligible; if any of these fail, the extracted $\\Gamma_{SD}$ is not the single-state scattering rate.","fun_headline_variants_meta":{"raw":{"variants":["Raman scattering rates confirm sub-1e-4 qubit gate errors","Metastable qubit scattering rates clear path to 1e-4 gates","Scattering out of Ca+ qubit measured, gate infidelity below 1e-4","Calcium ion qubit scattering rates enable ultra-low error gates","Ion qubit scattering measured: sub-1e-4 gate errors feasible"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000262,"raw_usage":{"total_tokens":1615,"prompt_tokens":983,"completion_tokens":632,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":599,"completion_tokens_details":{"reasoning_tokens":528}},"tokens_in":599,"tokens_out":632,"duration_ms":5671,"temperature":1.0,"reasoning_tokens":528,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T23:19:58.619708+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Shine $\\sigma^+$-polarized 976 nm light on the $|\\uparrow\\rangle$ state: the model says no dipole-allowed transition to $P_{3/2}$, so any measured scattering rate above noise would falsify the claim. A second check is to measure the $D_{5/2}$ decay curve at several intensities well beyond 1 s or with sublevel-resolved readout; a visible deviation from single-exponential decay or a rate that changes with intensity would indicate optical pumping or multiple-scattering effects the analysis neglects.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the parameter-free scattering-rate model for metastable $D_{5/2}$ qubits that the measured rates are compared against."},{"cited_title":"Ozeri, W","cited_arxiv_id":null,"evidence_quote":"Provides the spontaneous-scattering gate-error formula used to convert measured rates into single-qubit error probabilities."},{"cited_title":"Kreuter, C","cited_arxiv_id":null,"evidence_quote":"Supplies the natural $D_{5/2}$ lifetime used to extract the induced scattering rate from the measured decay constant."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the $P_{3/2}$ lifetime whose uncertainty dominates the theoretical error bars on the predicted rates."},{"cited_title":"Quinn, G","cited_arxiv_id":null,"evidence_quote":"Establishes the 976 nm Raman gate platform in calcium whose error floor is being characterized."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Prior verification of the theory for ground-state qubits in barium, motivating the metastable-manifold test performed here."},{"cited_title":"Song, Y.-B","cited_arxiv_id":null,"evidence_quote":"Supplies the $P_{3/2}$ branching fractions used to convert measured rates into scattering rates back to the $D_{5/2}$ manifold."}],"review_version":1}