{"id":"8dd64fb8-ab02-4ea9-9828-cda5a35a688e","arxiv_id":"2502.02398","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":2.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":5,"one_line_summary":"Lattice simulations of global axion string networks, corrected for discretization and oscillation effects, are extrapolated to predict the axion dark matter mass in the 95 to 450 micro-eV band.","lead":"This proceedings paper reports a numerical prediction that QCD axion dark matter from post-inflationary cosmic strings has a mass between 95 and 450 micro-eV. The result narrows the target frequency range for next-generation axion haloscope experiments, which need to know where to search.","discovery_kind":"review","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The predicted mass range depends on extrapolating q(ℓ) from ℓ≈9 to ℓ≈70 with four ad hoc fit forms (Eq. 11); none is derived, and the mass integral is sensitive to the resulting q.","rationale":"The reader's weakest assumption correctly identifies the extrapolation of q. I agree: this is the load-bearing step. I checked the paper for internal consistency: the equations are well-defined, the simulation code is public (jaxions), and the paper is transparent about systematic effects (initial conditions, oscillations, discretisation). There is no sign of circularity or an omitted proof; the issue is that the functional forms in Eq. (11) are unconstrained by theory. The claim is conditional by design—the conclusion is an envelope over four fits. My stress-test therefore does not change the reader's CONDITIONAL verdict. I considered raising the dependence on the string density ξ(ℓ) or the removal of spectrum oscillations, but those are secondary: even if ξ and the oscillation cleaning were perfect, the q extrapolation alone would control the final mass band. The concrete test via AMR is the most direct way to settle the question, and the authors themselves point to AMR as the next step.","tokens_in":9225,"tokens_out":3674,"duration_ms":37862,"concrete_test":"Run or use AMR simulations (as in Refs. [14,17]) to measure q(ℓ) up to ℓ ≈ 12–15 with m_r a < 0.5 throughout, apply the same oscillation/decomposition analysis, and compare the production efficiency K(ℓ) (left panel of Fig. 3) with the predictions of models A–D extrapolated to that ℓ. If the AMR point falls outside the band of the four models or shows a different trend (e.g., flattening toward q = 1), the quoted mass range in Eq. (14) cannot be trusted; if it falls inside, the conditional concern is resolved.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim (Eq. 14) rests on the extrapolation of the spectral index q(ℓ) to the cosmological value ℓ = ln(f_a/H) ≈ 70. The four models in Eq. (11) are empirical ansaetze (linear, quadratic, inverse, inverse-square in ℓ) that all fit the simulated range ℓ ≈ 3–9, but they diverge widely by ℓ = 70. Since the axion number density in Eq. (12) depends on the spectral shape F ~ x^{-q}, and the production efficiency K is strongly sensitive to the value of q over the integration range, the final mass band 95–450 μeV is essentially the spread of these arbitrary fitting forms, plus the fitted discretisation correction (10). The discretisation correction itself contains an exponential in ℓ and is extracted at m_r a values that approach and exceed 1 near the end of the runs, so its removal at physical m_r a -> 0 is not directly demonstrated. Section 6 concedes that adaptive mesh refinement is a 'possible next step', which is an explicit acknowledgement that the extrapolation is not yet independently confirmed. Without an independent handle on q at large ℓ, Eq. (14) is an envelope of model choices rather than a derived prediction.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This proceedings paper reports large-scale static-lattice simulations of global axion strings and uses them to predict the QCD axion dark-matter mass. The authors model the string-density attractor (Eqs. 3-4), define the instantaneous axion emission spectrum F(x) (Eqs. 6-9), and fit its spectral index q as q_model(ℓ)+q_disc(ℓ,m_r a) (Eq. 10) with four empirical late-time ansätze (Eq. 11). Combining these with the production efficiency K (Eq. 13) and the observed dark-matter abundance, they obtain the mass range 95 μeV ≲ m_a ≲ 450 μeV (Eq. 14), which overlaps next-generation haloscope sensitivity. The paper also discusses systematic effects—initial conditions, spectral oscillations, and discretisation—that may explain discrepancies with earlier work.","tokens_in":9790,"tokens_out":5424,"duration_ms":55842,"significance":"If correct, this prediction would focus experimental axion searches on a well-motivated μeV band. The paper deserves credit for a large numerical campaign, for identifying and quantifying the string-density attractor, for explicitly isolating discretisation and oscillation contaminations, and for propagating uncertainties in q, ξ, x0, and n_QCD into the final band (Fig. 3). However, the quoted range is obtained from fitted functional forms that are extrapolated far beyond the simulated dynamical range; it is a model-dependent extrapolation rather than a first-principles derivation. The value of the paper lies in the systematic treatment and in making the extrapolation transparent, not in a proof of Eq. (14) as it stands.","major_comments":[{"comment":"The central mass prediction is the envelope of empirical forms for q_model fitted at ℓ≈3–9 and extended to ℓ≈70. Eq. (11) lists four ansätze, yet Fig. 3 reports the mass extrapolation only for two of them (q0+q1ℓ and q0+q1/ℓ², each with ξlin/ξsat); it is unclear whether models B and C are used at all and how the quoted 95–450 μeV spread was selected. Because K in Eq. (12) is strongly sensitive to q through the spectrum F∼x^{-q}, and because the text itself identifies AMR as a 'possible next step' (Sec. 6), Eq. (14) should be reframed as the spread of model choices, or supported by an independent handle on q at large ℓ.","section":"§4, Eq. (11); §6, Eq. (14)"},{"comment":"The discretisation correction q_disc(ℓ,m_r a) has an exponential form whose fitted constants d_i are not given, and the simulations reach values m_r a ≳ 1 near the end of the runs while the physical limit is m_r a → 0. The text admits that the final spectra are visibly distorted by discretisation, so it is not demonstrated that the exponential fit, rather than real physical evolution of q, is absorbing the late-time rise. The authors should show stability of the q_model extrapolation under variation of lattice spacing at fixed ℓ, or cross-check with AMR or PRS-type simulations, before Eq. (14) can be considered robust.","section":"§4, Eq. (10)"},{"comment":"The attractor extrapolation itself contributes a factor-of-two spread: ξlin(70)≈13.8(5) versus ξsat(70)≈7(3). Figure 3 appears to include both ξ models, but the paper does not state whether the final band in Eq. (14) is a combined envelope over all q models and ξ models or only over selected combinations. Since K and m_a depend directly on the string density, the selection criterion for the models entering Eq. (14) should be explicit.","section":"§3, Eq. (5); §5, Fig. 3"}],"minor_comments":[{"comment":"There are several typos: \"apparant\" should be \"apparent\", \"instanteneous\" should be \"instantaneous\", \"casted\" should be \"cast\", and \"proclaimend\" should be \"proclaimed\".","section":"§2"},{"comment":"The legend in the right panel of Fig. 2 appears to show only three of the four q_model curves listed in Eq. (11); the q0+q1ℓ² model is missing from the legend, which makes it harder to see which models enter the final analysis.","section":"§4, Fig. 2"},{"comment":"Equation (12) uses τ and τ′ without defining the integration domain or the origin of the τ-dependent prefactor; a reader unfamiliar with Ref. [1] cannot reproduce the integral from this text alone.","section":"§5, Eq. (12)"},{"comment":"The uncertainty bands in Fig. 3 are labeled by source (q, ξ, x0, n_QCD), but the text does not explain how these sources are combined into the final quoted range in Eq. (14).","section":"§5, Fig. 3"},{"comment":"The statement that the simulation code is available on GitHub is helpful, but the paper does not specify the version or provide a reproducibility record; if space permits, a simulation metadata table (lattice size, m_r a range, run times) would strengthen the presentation.","section":"§2"}],"recommendation":"major_revision","confidential_remarks":null},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dear colleague,\n\nThe short take: this is a conference proceedings that repackages the authors' JCAP 2024 paper [1] into a readable summary. The headline result, 95 ≲ m_a ≲ 450 μeV, is credible as a statement of where their simulations land, but it is not a sharp prediction—it is the spread of four empirical extrapolation models for the spectral index q(ℓ). If you use it to plan a haloscope search, take the band as a target range, not a confidence interval.\n\nWhat the paper does well: it explains the main systematic effects—initial-condition dependence, oscillation contamination in the spectrum, and discretisation bias—with enough clarity to show why the older literature spans orders of magnitude. The updated comparison figure (Fig. 4) is genuinely useful, and the discussion of the string-density attractor is a nice pedagogical touch. The authors are transparent about the extrapolation's limitations, even pointing to AMR as the obvious next step.\n\nThe soft spots are exactly where the stress-test note lands. The mass integral (Eq. 12) is sensitive to q(ℓ), and the four fit forms in Eq. (11) all describe the simulated range ℓ≈3–9 but diverge by ℓ=70. Since those forms are not derived, the final band is partly a measure of model choice. The discretisation correction in Eq. (10) is extracted at m_r a values that reach and exceed unity near the end of the runs, so its removal to the continuum limit (m_r a→0) is an extrapolation layered on top of an extrapolation. The paper doesn't hide this, but a reader who misses the fine print could over-interpret the mass range.\n\nOne more thing: as a standalone preprint, it contains no new derivations or data beyond the companion paper. The key technical details are in Ref [1], so don't cite this proceedings for the result; cite the JCAP paper.\n\nBottom line: this is a fine entry point for experimentalists and newcomers, and it's an honest summary of careful work. But a serious referee would treat it as a proceedings, not as a primary claim. I'd recommend it for the record, with no more than a light editorial pass, and point readers to Ref [1] for the full method.","headline":"Honest proceedings summary of a solid simulation campaign; the headline mass range is an envelope of extrapolated fits, not a sharp prediction.","tokens_in":10085,"tokens_out":3783,"would_cite":false,"duration_ms":37126,"reading_group":"maybe","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 predicts the QCD axion dark matter mass at 95–450 μeV from the spectrum of axions radiated by global cosmic strings, and argues this range will be probed by next-generation haloscopes.","keywords":["QCD axion","axion dark matter","global cosmic strings","cosmic string network","spectral index","Peccei-Quinn symmetry breaking","lattice simulation","haloscope searches"],"falsifier":"Run a sufficiently large adaptive-mesh-refinement simulation that reaches $\\ln(m_r/H) \\gtrsim 12$ with controlled core resolution and measure $q(\\ell)$ directly; if the measured late-time spectral index deviates from all four model extrapolations by more than the stated uncertainty, the mass band in Eq. (14) shifts. A complementary check is an axion search with sufficient sensitivity that finds no axion across the entire 95–450 $\\mu$eV range, which would contradict the claim that string-produced axions are all of the dark matter.","tokens_in":9032,"feed_emoji":"🌀","tokens_out":12497,"duration_ms":107289,"temperature":0.7,"pith_summary":"In the post-inflationary Peccei-Quinn scenario—where the axion's symmetry breaks after inflation and necessarily leaves a network of global cosmic strings—the axions radiated when strings decay would today be the dark matter. This paper tries to make that production channel yield a sharp mass prediction by measuring the axion emission spectrum in large numerical simulations. It catalogs systematic effects that earlier work under-reported: dependence on initial conditions, contamination from oscillating fields in the spectrum, and lattice discretisation, all of which bias the extracted spectral index. After correcting for these and extrapolating to the cosmological string tension, it predicts the axion dark matter mass lies between $95\\,\\mu\\text{eV}$ and $450\\,\\mu\\text{eV}$ (Eq. 14). If right, the QCD axion in this scenario sits directly in the band next-generation haloscopes are being built to scan.","feed_headline":"Axion dark matter mass predicted at 95–450 μeV","feed_subtitle":"Cosmic string decay sets the QCD axion mass band that next-generation haloscopes will scan.","key_machinery":"The carrying object is the dimensionless instantaneous axion emission spectrum $F(x,y)$ defined in Eq. (6), together with its spectral index $q$ from the power-law ansatz of Eq. (9). The decisive mechanism is the split in Eq. (10), $q = q_{\\rm model}(\\ell) + q_{\\rm disc}(\\ell, m_r a)$, which separates the physical late-time evolution from the lattice-core artifact controlled by $m_r a$; the four fitted models A–D in Eq. (11) are how the paper handles the extrapolation in $\\ell$. The string density enters through the attractor equation $d\\xi/dt = (C/t)(\\xi_c(\\ell) - \\xi)$, calibrated on conformal networks, which determines how many strings radiate. These ingredients feed the axion number density (Eq. 12), the production-efficiency factor $K = n^{\\rm str}_a/n^{\\rm mis}_a$, and the matching condition $\\Omega_a h^2 = K \\Omega^{\\rm mis}_a h^2 = 0.12$ that fixes $m_a$.","core_discovery":"The paper's central claim is that the axion mass produced by global string decay is $95\\,\\mu\\text{eV} \\lesssim m_a \\lesssim 450\\,\\mu\\text{eV}$. The argument proceeds through the instantaneous emission spectrum $F(x,y)$, modeled as a power law $F = F_0 x^{-q}$ on $x_0 < x < y$, with $q$ the spectral index that decides whether the axion population is soft-dominated or hard-dominated. The paper decomposes $q$ into a late-time physical model $q_{\\rm model}(\\ell)$ plus a discretisation correction $q_{\\rm disc}(\\ell, m_r a)$ (Eq. 10), fits four functional forms (A–D) for $q_{\\rm model}$, and confirms an attractor evolution for the string density $\\xi$ whose extrapolation to $\\ell = \\ln(f_a/H) \\approx 70$ gives $\\xi \\sim 7$ to $13.8$. Feeding these into the number-density integral (Eq. 12) and matching $\\Omega_a h^2 = 0.12$ yields the mass band in Eq. (14).","pith_inferences":["A purely numerical test of the extrapolation is available now: rerun the same physical initial conditions with adaptive mesh refinement to push $\\ln(m_r/H)$ toward 12 or beyond and see whether the directly measured $q(\\ell)$ follows any of the four models A–D; if the best late-time model changes, the 95–450 $\\mu$eV band should shift.","The same $\\ell$-extrapolation logic applies to other global topological defects whose energy diverges logarithmically, such as global monopoles and domain walls, so the correction scheme may carry over to their emission spectra.","A detection inside this band would not by itself prove the string-production picture, but a null result across the entire 95–450 $\\mu$eV band would make it hard to sustain the post-inflationary PQ explanation of all the dark matter under the stated cosmological assumptions.","The claim that $\\xi$ keeps growing to roughly 7–14 at $\\ell=70$ extends the conformal-network attractor to parametrically large $\\ell$; simulations with much larger scale separation could confirm or break that extension."],"forward_implications":["The post-inflationary PQ axion becomes discoverable in a defined decade-wide mass window: experiments such as FLASH, IAXO, RADES, ADMX, and QUAX can cover the lower half of the 95–450 $\\mu$eV range, while ALPHA, MADMAX, and ORGAN target the upper half.","The predicted range gives haloscope searches a concrete frequency band of roughly 23–109 GHz, so experimental programs can prioritize cavity designs and magnet volumes for that band.","Earlier simulations spread over a few to roughly 1000 $\\mu$eV; the systematic catalog here attributes that spread to identifiable biases, implying future lattice studies that control the same effects should converge to a common spectral index.","If the extrapolated string density $\\xi \\sim 7$ to $13.8$ at $\\ell=70$ is correct, string decay contributes substantially more axions than the misalignment mechanism, making the string channel the dominant source of axion dark matter in this scenario."],"supporting_citations":[{"why":"It is the companion paper whose spectrum analysis, fits, and figures this contribution summarizes.","marker":"[1]"},{"why":"It introduced the attractor solution for the string density that this work confirms and models.","marker":"[12]"},{"why":"It provides an earlier determination of the spectral index and axion abundance that this work compares against and revises.","marker":"[13]"},{"why":"It provides the adaptive-mesh-refinement simulation with a larger dynamical range and a near-conformal spectrum, the main competing benchmark.","marker":"[14]"},{"why":"It supplies the conformal string network data used to calibrate the two attractor models for $\\xi_c(\\ell)$.","marker":"[21]"},{"why":"It introduces the discretisation parameter $m_r a$ and the statement that coarse string cores bias the spectral index upward.","marker":"[24]"},{"why":"It provides the high-temperature lattice QCD calculation entering the misalignment normalization that sets the scale of the mass prediction.","marker":"[26]"},{"why":"It supplies the precise QCD axion parameters used for the misalignment estimate that defines the efficiency factor $K$.","marker":"[27]"},{"why":"It gives the observed dark matter abundance $\\Omega_{\\rm CDM}h^2 \\approx 0.12$ used as the matching condition.","marker":"[28]"}],"fun_headline_variants":["Axion mass 95–450 μeV from cosmic string decay","String decay narrows axion dark matter mass to 95–450 μeV","Global string spectrum sets axion dark matter mass: 95–450 μeV","95–450 μeV: Axion dark matter mass from string decay"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the four fitted models for how the spectral index $q(\\ell)$ changes at late times, together with the discretisation correction $q_{\\rm disc}(\\ell, m_r a)$, remain valid when extrapolated from the simulated range $\\ln(m_r/H)\\approx 3$ to 9 out to the cosmological value $\\ell = \\ln(f_a/H)\\approx 70$; a different bend in $q$ outside the simulated range moves the mass band in Eq. (14).","fun_headline_variants_meta":{"raw":{"variants":["Axion mass 95–450 μeV from cosmic string decay","String decay narrows axion dark matter mass to 95–450 μeV","Global string spectrum sets axion dark matter mass: 95–450 μeV","95–450 μeV: Axion dark matter mass from string decay"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001184,"raw_usage":{"total_tokens":4909,"prompt_tokens":987,"completion_tokens":3922,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":603,"completion_tokens_details":{"reasoning_tokens":3840}},"tokens_in":603,"tokens_out":3922,"duration_ms":24557,"temperature":1.0,"reasoning_tokens":3840,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-09T12:17:06.099398+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Run a sufficiently large adaptive-mesh-refinement simulation that reaches $\\ln(m_r/H) \\gtrsim 12$ with controlled core resolution and measure $q(\\ell)$ directly; if the measured late-time spectral index deviates from all four model extrapolations by more than the stated uncertainty, the mass band in Eq. (14) shifts. A complementary check is an axion search with sufficient sensitivity that finds no axion across the entire 95–450 $\\mu$eV range, which would contradict the claim that string-produced axions are all of the dark matter.","supporting_citations":[{"cited_title":"Global cosmic string networks as a function of tension","cited_arxiv_id":"1912.08058","evidence_quote":"It supplies the conformal string network data used to calibrate the two attractor models for $\\xi_c(\\ell)$."}],"review_version":1}