REVIEW 2 major objections 3 minor 6 references
Comment on arXiv:1901.10843 "Search for Axionlike Dark Matter with a Liquid-State Nuclear Spin Comagnetometer"
T0 review · 2 major / 3 minor · reviewed 2026-08-14 · deepseek-v4-flash
Pith's one-line read A 2019 axion-dark-matter limit is too strong by a factor of 9.5, a comment argues.
desk verdict Real flaw found in Wu et al.'s long-period axion limit, but the corrected factor is asserted, not derived. read the letter →
The pith
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
The reading
What carries the argument
The load-bearing tool is the linear least-squares analysis (LLSA) of the time series, with two basis functions per assumed axion mass: one sine and one cosine at the axion frequency, transformed from celestial to laboratory coordinates. The LLSA reports the two quadrature amplitudes and their correlation; the crucial input is how that correlation behaves when the oscillation period exceeds the data span. The identity that carries the argument is coupling $\propto \Delta R/\sin\varphi$, which shows that the unknown phase must be treated as a free parameter rather than averaged out.
What would settle it
Reanalyze the original $2.642\times10^6$ s time series using the linear least-squares fit described here, leaving the axion phase free, and read the upper limit at the left-hand edge of the period range; if the limit moves up by less than a factor of about 9.5 relative to the published curve, the comment's central number is wrong. A cheaper check is to simulate a pure-noise data set with the same duration and sampling and confirm that the marginalized amplitude error scales as $\tau/(\pi T)$ once $\tau>T$.
Extended reading notes
Core claim
On the paper's own terms, the original experiment's published constraint is not valid for axion oscillations whose period $\tau$ is comparable to or longer than the data span $T = 2.642\times10^6$ s. The correct treatment is a linear least-squares fit of the time series to two quadrature basis functions; when $\tau/T>1$, these basis functions become nearly parallel or anti-correlated, so the marginalized amplitude uncertainty grows as $\tau/(\pi T)$. At the left-hand edge of the original Fig. 3 this growth is a factor of 6.0, and the supplementary analysis's phase averaging adds a further error because the coupling scales as $1/\sin\varphi$, not as the average of $|\sin\varphi|$. The combination means the low-mass exclusion limit should be 9.5 times higher than plotted, leaving the neutron-EDM bound about 15% more constraining.
Load-bearing premise
The argument assumes the unknown axion phase should be fitted as a free parameter, with uncertainty scaling as $1/\sin\varphi$; if the original paper's phase-averaging were a valid statistical construction, the factor 9.5 would not follow.
Editorial extensions
If this is right
- The published low-mass edge of the original exclusion curve is not valid as stated.
- A least-squares reanalysis would move that edge upward by a factor of 9.5.
- The corrected limit is about 15% less constraining than the neutron-EDM limit, which therefore remains the strongest bound in this mass range.
- The same sensitivity degradation applies to any oscillation search whose period exceeds the observation span, so finite run duration must appear explicitly in the error budget.
- The short-period side of the original curve is assumed correct and unaffected.
Reading between the lines
- If the comment is right, other axion-dark-matter searches that marginalized over the axion phase by averaging $|\sin\varphi|$ may need to revisit their long-period limits with a free-phase fit.
- A testable extension: simulate the original run's sampling and noise, fit with LLSA for a grid of $\tau/T$ values, and verify that the limit tracks $1/\sin\varphi$ and $\tau/(\pi T)$ in the long-period regime.
- The comment's logic implies that for periods beyond the run length, extending the data span improves sensitivity faster than reducing per-measurement noise, a point the authors leave implicit.
Signed reviews
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The manuscript is a comment on Wu et al. (arXiv:1901.10843), which reported a search for axionlike dark matter using a liquid-state nuclear spin comagnetometer. The comment argues that the long-period part of the published exclusion limit (left panel of Fig. 3) is invalid because the data span T=2.642×10^6 s is much shorter than the oscillation periods probed at the left edge. It proposes a linear least-squares analysis (LLSA) in which the sine and cosine quadrature basis functions become nearly degenerate for τ>T, causing the sine amplitude uncertainty to grow as τ/(πT); at the left edge of the published figure this factor is 6.0. It further criticizes the original supplementary-material phase treatment (Eqs. S12-S17) for averaging |sin φ| when the coupling-constraint expression contains 1/sin φ. Assuming the short-period analysis is correct, the comment concludes that the left-edge exclusion limit should be 9.5 times weaker, making it about 15% less constraining than the previous neutron-EDM limit.
Significance. If the arguments are correct, the comment identifies a genuine flaw in a published experimental upper limit and provides a simple, parameter-free explanation for the degradation of sensitivity at long periods. The LLSA framework is a useful cross-check, and the criticism of phase-averaging is statistically well founded. The main quantitative conclusion, however, depends on the factor 9.5, which is not derived in the manuscript; before the comparison to the neutron-EDM limit can be accepted, the derivation must be supplied. The paper otherwise introduces no free parameters and is consistent with other independent searches.
major comments (2)
- [Final paragraph (9.5 factor)] The factor 9.5 in the final paragraph is asserted without derivation. The text establishes a degradation factor τ/(πT)=6.0 for the sine amplitude at the left edge of Fig. 3 and separately argues that the original phase-averaging of |sin φ| is incorrect because the constraint scales as 1/sin φ, but it never shows how these two effects combine to produce 9.5. Since the concluding comparison to the neutron-EDM limit (∼15% less constraining) rests entirely on this number, the authors should state the explicit calculation, e.g., the product of 6.0 with the phase-correction factor and any additional normalization factors, and justify each step with reference to the original analysis equations.
- [LLSA example] The claim that the extracted sine amplitude error 'blow[s] up as τ/(πT)' is stated after a short Taylor expansion, but the derivation of the factor is not shown. In particular, the normalization of the basis functions and the construction of the marginalized amplitude uncertainty from the correlation matrix are omitted. A concise derivation (or a reference to the analogous calculation in ref. [5]) would make the factor 6.0 verifiable and would support the subsequent 9.5 factor.
minor comments (3)
- [Paragraph on 'strange long-period behavior'] The phrase 'the constraint actually becomes tighter as πT/τ < 2.5' is confusing; the inequality should be expressed in terms of ξ=τ/T for consistency with the rest of the text.
- [General] The comment would benefit from reproducing or referencing the specific panel of Fig. 3 that is criticized, since the 'left-hand edge' is not identifiable without the original paper at hand.
- [Typo] The sentence 'Note that approaches where the φ ia a free parameter will not have a pole for any non-zero frequency' contains a typo: 'ia' should be 'is'.
Circularity Check
No circularity: the comment is an independent statistical critique whose central factor is derived from the target paper's equations and standard least-squares reasoning, not from its own conclusions.
full rationale
The paper is a comment on Wu et al. (arXiv:1901.10843). Its central claim — that long-period axion constraints scale as tau/(pi T) and that the target paper's phase-averaging prescription is incorrect — is derived from standard linear least-squares orthogonality arguments and from the target paper's own Eqs. S12-S17, not from the commenters' prior results. No parameters are fitted to the target data, and the correction factor 9.5 is presented as a consequence of combining the 6.0 blow-up with the target paper's sin(phi) relation. The only self-citation is to ref. [5] (Terrano et al.), invoked as an example of a linear least-squares analysis; the comment independently explains the basis functions and does not make the argument depend on that citation. A separate transparency concern, not a circularity, is that the arithmetic producing 9.5 from the stated 6.0 factor is not displayed and the phase-averaging correction is not quantified in the text; this could affect the headline number but does not make the derivation depend on its own conclusion. The comment is self-contained against the target paper's equations and external constraints.
Assumptions & free parameters
assumptions (3)
- domain assumption The data set duration T = 2.642 × 10^6 s is taken from the original paper (Wu et al.).
- domain assumption The original analysis for signals with τ < T is correct, so the error is confined to the long-period regime.
- domain assumption The axion signal is modeled as an oscillation with a single frequency and unknown phase φ, and the appropriate statistical treatment treats φ as a free parameter in a least-squares fit.
Cite this review
Pith. "Pith review of Comment on arXiv:1901.10843 "Search for Axionlike Dark Matter with a Liquid-State Nuclear Spin Comagnetometer"." pith.science (2026). https://pith.science/paper/MXK2B4P4
@misc{pith2026190810232,
author = {Pith},
title = {Pith review of: Comment on arXiv:1901.10843 "Search for Axionlike Dark Matter with a Liquid-State Nuclear Spin Comagnetometer"},
year = {2026},
howpublished = {\url{https://pith.science/paper/MXK2B4P4}},
note = {Machine review of arXiv:1908.10232}
}
read the original abstract
We comment on arXiv:1901.10843, pointing out that its constraint on ultra-low-mass axions is incompatible with the duration of the data set described. We describe a simple way to analyze such data that gives the correct scaling.
Reference graph
Works this paper leans on
- [5]
-
[1]
Teng Wu et al., Phys. Rev. Lett. 122, 191302 (2019)
work page 2019
-
[2]
S12-S17) in the supplementary material
5, to the discussion in Derivation of The Constraint Level (Eqs. S12-S17) in the supplementary material. These relate the data ∆ R to the axion coupling gaN N and the phase φ of the axion oscillation, with gaN N pro- portional to ∆ R/ sin φ. The authors account for the ef- fect on gaN N of the unknown phase by averaging |sin φ| over the interval for φ bet...
-
[3]
K. Van Tilburg, N. Leefer, L. Bougas, and D. Budker, Phys. Rev. Lett. 115 011802 (2015)
work page 2015
-
[4]
A. Hees, J. Gu´ ena, M. Abgrall, S. Bize, and P. Wolf, Phys. Rev. Lett. 117 061301 (2016)
work page 2016
- [6]
Reviewed August 14, 2026 · model on record in the stance chip above.
Discussion (0). Continue with ORCID to comment.