{"id":"ea4ae297-b6df-4ac5-8b26-0ff78abef6b5","arxiv_id":"2601.09115","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"SASHMAG retrieves 0.2–0.4 T magnetic fields from 87Rb saturated-absorption line centers with ±0.0017 T statistical precision, supported by a multilevel optical Bloch-equation model in the |mI,mJ> basis.","lead":"A rubidium vapor-cell sensor measures magnetic fields from 0.2 to 0.4 tesla by reading the positions of narrow saturated-absorption lines in the hyperfine Paschen-Back regime. The accompanying optical Bloch-equation model reproduces the sub-Doppler spectra and is positioned as a basis for future automated magnetometry.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Field-inference claim rests on unvalidated one-to-one mapping from Gaussian peak centers to Hamiltonian eigenfrequencies; overlapping peaks (e.g., σ− peaks 4/5 at 0.4092 T, separated by 6 MHz) and crossover resonances can bias centers by more than the claimed ±1.7 mT.","rationale":"I focused on the central quantitative claim: field retrieval precision. The reader's weakest assumption is exactly the one I find most load-bearing. The paper provides no point-by-point comparison of measured and simulated line shapes, no measurement of sub-Doppler linewidths, and no discussion of overlapping resonance or crossover effects. The only numeric error propagation is statistical. Given the 6 MHz spacing between peaks 4 and 5 in Table I, a modest linewidth (tens of MHz for room-temperature Rb with power broadening) would cause a Gaussian center bias well above 24 MHz, i.e., >1.7 mT. Thus the claimed precision may be overly optimistic. This is not a challenge to the atomic physics or to the OBE model in general, but to the inference step. I agree with the reader and see no need to change the CONDITIONAL verdict; the paper should add this systematic-bias check before stronger claims are accepted.","tokens_in":13080,"tokens_out":4731,"duration_ms":45067,"concrete_test":"Run the paper's Gaussian peak-extraction and field-optimization pipeline on a synthetic SAS spectrum generated by the OBE model at a known field B_true = 0.4092 T under the same experimental parameters (probe waist, pump power, temperature). Compare the inferred B to B_true. If the bias |B_inferred − B_true| exceeds 0.0017 T, the systematic error from overlapping lines is comparable to or larger than the reported precision, and the claim is not supported. Additionally, fit the measured spectrum at 0.4092 T with a two-Gaussian model for peaks 4 and 5 and check whether the fitted centers shift by more than ~24 MHz (the frequency equivalent of 1.7 mT at the local Zeeman slope) compared to a single-Gaussian fit.","verdict_should_be":"UNCHANGED","load_bearing_attack":"In Sec. IV C, the loss function Eq. (8) compares Gaussian-fitted experimental peak positions ν_exp^(k) with eigenfrequencies ν_theo^(k)(B) from diagonalizing Eq. (1). This assumes every fitted peak is an isolated Doppler-free resonance whose center equals a single transition frequency. That assumption is not tested. At B=0.4092 T, Table I lists σ− peaks 4 and 5 at −7898.86 and −7905.06 MHz, separated by only 6.2 MHz. If their sub-Doppler linewidths are comparable to this spacing (the paper reports no measured linewidths), the two features blend and a single Gaussian fit will return a biased center. Crossover resonances, known to appear in Rb saturation spectroscopy in magnetic fields (Ref. [27]), are neither identified nor excluded. The Monte Carlo error analysis (Eqs. 9–10) propagates only statistical fitting and calibration uncertainties; it does not include this systematic line-pulling bias. Model-vs-experiment agreement shown in Fig. 4 is only qualitative, so it cannot rule out line-shape discrepancies that would affect peak centers. The claimed precision of ±0.0017 T therefore rests on an assumption that has not been validated.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper presents SASHMAG, a rubidium-87 vapor-cell magnetometer operating in the hyperfine Paschen-Back regime at fields from 0.2 T to 0.4 T. The authors develop a multilevel optical Bloch-equation (OBE) model in the |m_I, m_J> basis to simulate saturated absorption spectra at sub-Doppler resolution, and they infer the magnetic field by minimizing the squared difference between experimentally extracted Gaussian peak centers and Hamiltonian eigenfrequencies. They report field estimates of 0.4092 ± 0.0017 T, 0.3263 ± 0.0016 T, and 0.2602 ± 0.0016 T, with a sensitivity of 0.42 mT/√Hz. The OBE model is validated against measured spectra qualitatively (Fig. 4) and against analytical power-broadening and Doppler-scaling predictions using simulated spectra (Fig. 5).","tokens_in":13443,"tokens_out":2056,"duration_ms":20140,"significance":"If the central claims hold, this work provides a useful route to absolute magnetic-field measurement in the intermediate-to-high field regime without thin-cell techniques or single-line tracking. The use of a multilevel OBE model in the uncoupled basis is a technically appropriate extension of earlier work by Maguire et al., and the inverse-problem approach of fitting Hamiltonian line positions to measured peak centers is physically well motivated. The Monte Carlo uncertainty propagation in Eqs. (9)-(10) is a sound way to estimate statistical precision, and the reported ±0.0017 T precision is plausible if the peak-center extraction is unbiased. However, the paper's central claim of a 'validated simulation' and the quoted precision rest on an untested assumption that each fitted Gaussian peak corresponds one-to-one to an isolated Doppler-free transition frequency. The absence of a quantitative model-vs-experiment comparison, and the presence of closely spaced peaks and potential crossover resonances, make the current evidence insufficient to support the stated precision and the claim that the OBE model reproduces measured spectra.","major_comments":[{"comment":"The sensitivity of 0.42 mT/√Hz is quoted without derivation or supporting data. No noise spectrum, bandwidth specification, or repeated-measurement analysis is provided. If this is intended as a sensor performance figure, it needs a clear definition (e.g., standard deviation of field estimates as a function of integration time) and experimental evidence. If it is derived from the Monte Carlo uncertainty and a single scan duration, that should be stated explicitly. Without this, the sensitivity claim is not assessable.","section":"Sec. IV C, 'sensitivity'"}],"minor_comments":[{"comment":"There is a typographical error in the abstract: 'with a precision of ±0.0017 T)' has an unmatched parenthesis. Also, 'we demonstrate magnetic field retrieval from 0.2 T to 0.4 T with a precision of ±0.0017 T' appears twice in slightly different forms in the abstract and introduction.","section":"Abstract"},{"comment":"The notation ρ_avg^{ββ}(v_i, Δ) is not formally defined before use. It is clear from context that it is the time-averaged excited-state population from the OBE, but a brief definition would improve readability.","section":"Sec. II C, Eq. (7)"},{"comment":"The caption of Fig. 2 describes the reference optics arm as 'top, orange section' and main experiment as 'bottom, blue section,' but the figure appears to be grayscale in the print version. If color is not used, the text should refer to spatial positions or use distinguishable labels.","section":"Sec. III, Fig. 2"},{"comment":"In Eq. (A24), the index in the term −i(ω_βα − ω)˜ρ_βγ appears to have a typo: it should be ˜ρ_βα, not ˜ρ_βγ. The subsequent derivation uses ˜ρ_βα, so this is likely a typographical error.","section":"Appendix A, Eq. (A24)"},{"comment":"Ref. [14] is a commercial spec page (Lake Shore) with no author or year; it would be better cited as a product manual with access date. Ref. [36] is a thesis; if used for the optimization algorithm, a more accessible reference or description in the text would be helpful.","section":"References"}],"recommendation":"major_revision","confidential_remarks":"The paper addresses an interesting and practically relevant problem, and the Hamiltonian-based inverse method is technically sound. However, the current validation is insufficient to support the central precision claim. The main concern is not that the approach is fundamentally flawed, but that the systematic uncertainties in peak-center extraction (overlapping peaks, crossover resonances, line-pulling) are not quantified. The authors should either provide a point-by-point model-vs-experiment comparison, demonstrate that fitted peak centers are unbiased at the claimed level, or explicitly reduce the claimed precision to reflect unquantified systematics. If this can be addressed, the paper could be acceptable; as it stands, the evidence does not support the stated claims."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Two things to know. The paper does something real: it extends the Maguire zero-field OBE framework to 87Rb in the hyperfine Paschen-Back regime, solving in the |mI,mJ> basis, and demonstrates a saturated-absorption magnetometer that works at 0.2–0.4 T. That combination is new and genuinely useful for MRI and fusion diagnostics. The OBE model captures the essential physics, and the authors are appropriately transparent about its limits—no collisional broadening, no number-density scaling, normalized amplitudes. The field inference itself is legitimate inverse problem-solving, not circular. The weak link is the evidence chain for the headline precision. The claim that each Gaussian-fitted experimental peak center maps one-to-one to a Hamiltonian eigenfrequency is never tested. At B=0.4092 T, the σ− peaks 4 and 5 are separated by only 6 MHz. If the experimental sub-Doppler linewidth is comparable to that, the two features overlap and the fitted center will be pulled. Crossover resonances, which the authors cite in Ref. [27], are not identified or excluded. The power-broadening and Doppler-scaling checks are run on simulated spectra, so they validate the model's internal consistency but not its agreement with the measured lineshape. The Monte Carlo error analysis propagates only statistical fitting and calibration uncertainties, leaving out any systematic line-pulling bias. And the sensitivity figure, 0.42 mT/√Hz, appears without a stated measurement bandwidth. None of this is fatal—the concept is sound and the model is a reasonable forward tool. But the ±0.0017 T precision is not yet supported until the lineshape assumption is checked. The fix is straightforward: compare the OBE lineshape to the measured spectrum point-by-point at one field, and report measured linewidths. The paper is for people working on high-field atomic magnetometry or using OBE for saturated-absorption modeling. It deserves a serious referee, though the revision should require that comparison and a defined sensitivity measurement. I'd give it conditional acceptance with revision.","headline":"Solid OBE extension and a working 0.2–0.4 T rubidium magnetometer, but the headline ±0.0017 T precision rests on an untested one-to-one mapping from Gaussian peak fits to Hamiltonian eigenfrequencies.","tokens_in":13940,"tokens_out":1734,"would_cite":false,"duration_ms":19709,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["32.60.+i","07.55.Ge"],"model":"deepseek-v4-flash","headline":"A multilevel optical Bloch-equation model in the uncoupled |mI, mJ> basis reproduces the sub-Doppler saturated-absorption spectra of 87Rb in the hyperfine Paschen-Back regime and retrieves magnetic fields from 0.2 to 0.4 T with a precision","keywords":["saturated absorption spectroscopy","rubidium-87 magnetometry","hyperfine Paschen-Back regime","optical Bloch equations","high-field magnetometry","Doppler-free spectroscopy","Zeeman transitions","physics-constrained optimization"],"falsifier":"A point-by-point comparison of the OBE-simulated lineshape with the measured spectrum at a field where two transitions approach crossing (e.g., near B where the σ− lines 4 and 5 become nearly degenerate) would reveal whether the Gaussian-fitted centers deviate from the Hamiltonian eigenfrequencies beyond the claimed ±0.0017 T equivalent. If the centers shift without a corresponding change in the Hamiltonian eigenvalues, the central assumption fails.","tokens_in":13034,"feed_emoji":"🧲","tokens_out":4650,"duration_ms":41831,"temperature":0.7,"pith_summary":"This paper claims that a rubidium-87 vapor cell, interrogated with counter-propagating pump and probe beams in a strong magnetic field, can serve as an absolute magnetometer in the hyperfine Paschen-Back regime (0.2–0.4 T). The key is a multilevel optical Bloch-equation model solved in the uncoupled |mI, mJ> basis, which reproduces the measured sub-Doppler saturated-absorption spectra, including state mixing and saturation broadening. Using this model to compute the field-dependent transition frequencies, the authors infer magnetic fields from the positions of the Doppler-free Zeeman peaks with ±0.0017 T precision and 0.42 mT/√Hz sensitivity. If correct, this provides a compact, calibration-free route to high-field magnetometry without thin cells or single-line tracking, and a validated forward model for generating synthetic training data for machine-learning sensors.","feed_headline":"Rubidium vapor cell reads 0.2–0.4 T fields to 0.0017 T","feed_subtitle":"A multilevel Bloch model turns Zeeman spectra into a physics-based gauge for MRI and fusion.","key_machinery":"The key machinery is the multilevel optical Bloch-equation model solved in the uncoupled |mI, mJ> basis. Starting from the Liouville–von Neumann equation with a magnetic-field-dependent Hamiltonian, the authors derive a set of 40 coupled differential equations for the populations and coherences of the 24 sublevels of the 87Rb D2 line (8 ground, 16 excited), after applying the rotating-wave approximation and selection rules. The model incorporates power broadening via the Rabi frequency and Doppler averaging over the Maxwell–Boltzmann velocity distribution. The resulting fluorescence signal is compared to experiment. The field-inference step then uses the same Hamiltonian's eigenfrequencies a","core_discovery":"The central discovery is that the multilevel optical Bloch-equation framework, previously used for zero-field saturated absorption, can be transplanted to the hyperfine Paschen-Back regime of 87Rb by re-expressing the density-matrix equations in the uncoupled |mI, mJ> basis. In this basis the field-dependent Hamiltonian is diagonalized directly, and the 24 sublevels of the D2 line are connected by σ+ and σ− transitions. The model reproduces the measured spectra at B = 0.4092 T (and other fields), with power-broadening ratios and Doppler-width scaling matching the analytic two-level predictions. The authors then treat the spectrum as a fingerprint: they optimize the single parameter B to mini","pith_inferences":["The one-to-one mapping between Gaussian-fitted line centers and Hamiltonian eigenfrequencies could be extended to a full-lineshape fitting that uses the OBE spectrum directly, potentially removing biases from unresolved crossover resonances and overlapping manifolds.","Because the eigenfrequencies in the |mI, mJ> basis are nearly linear in B at these fields, the inverse problem may reduce to a simple linear regression over peak shifts, making the sensor even easier to deploy in dynamic environments.","The same modeling approach could be applied to other alkalis (e.g., 85Rb, 133Cs) or to the D1 line, widening the field range and allowing cross-checks by probing different manifolds.","A direct test of the assumption could be to scan the magnetic field continuously and verify that the fitted peak centers track the Hamiltonian predictions to within the claimed precision across intervening crossings, where crossover resonances are expected."],"forward_implications":["Magnetic fields between 0.2 and 0.4 T can be read from a single saturated-absorption spectrum with sub-mT precision, without thin-cell geometries or tracking a single extreme-state resonance.","The validated OBE model generates high-fidelity synthetic spectra, which can train machine-learning algorithms for autonomous, real-time field estimation in MRI and fusion environments.","The method separates the problem into a physics-based forward model and a scalar parameter fit, so the same sensor hardware can be re-calibrated for different field ranges simply by changing the Hamiltonian.","The demonstrated power-broadening and Doppler-scaling agreement indicates the model captures the dominant nonlinear physics, making it a candidate replacement for linear-susceptibility codes in Doppler-free high-field spectroscopy."],"fun_headline_variants":["Rubidium magnetometer hits 0.0017 T precision in high fields","Bloch-equation model sharpens Rb magnetometry to 0.2–0.4 T","SASHMAG: Rb sensor measures 0.2–0.4 T with 0.0017 T uncertainty","Rb magnetometer uses Bloch model for 0.2–0.4 T field sensing"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The field-inference pipeline assumes each measured saturated-absorption peak is an isolated Doppler-free resonance whose Gaussian-fitted center maps one-to-one to an eigenfrequency of the field-dependent Hamiltonian; unresolved crossover or overlapping σ+ and σ− lines could bias the retrieved field.","fun_headline_variants_meta":{"raw":{"variants":["Rubidium magnetometer hits 0.0017 T precision in high fields","Bloch-equation model sharpens Rb magnetometry to 0.2–0.4 T","SASHMAG: Rb sensor measures 0.2–0.4 T with 0.0017 T uncertainty","Rb magnetometer uses Bloch model for 0.2–0.4 T field sensing"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001112,"raw_usage":{"total_tokens":4500,"prompt_tokens":808,"completion_tokens":3692,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":552,"completion_tokens_details":{"reasoning_tokens":3600}},"tokens_in":552,"tokens_out":3692,"duration_ms":24319,"temperature":1.0,"reasoning_tokens":3600,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-03T10:41:09.344451+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A point-by-point comparison of the OBE-simulated lineshape with the measured spectrum at a field where two transitions approach crossing (e.g., near B where the σ− lines 4 and 5 become nearly degenerate) would reveal whether the Gaussian-fitted centers deviate from the Hamiltonian eigenfrequencies beyond the claimed ±0.0017 T equivalent. If the centers shift without a corresponding change in the Hamiltonian eigenvalues, the central assumption fails.","supporting_citations":[],"review_version":1}