{"id":"20e8d367-0f6d-406d-9742-1aa7790dfd6a","arxiv_id":"2502.08971","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":5,"one_line_summary":"Strong X-ray emission from Population III stars leaves a detectable enhancement in the 21cm power spectrum at z=7-10, provided their X-ray luminosity per star formation is at least an order of magnitude above Population II.","lead":"This paper models Population III star formation in a large cosmological simulation and computes its imprint on the 21cm hydrogen signal. It finds that strong X-ray emission from these first stars heats the early universe and could be detectable by SKA1-low at redshifts 7 to 10.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The SKA detectability claim rests on an unconstrained LX/SFR for Pop. III; the paper's own weak model shows D<1, and the moderate/extreme values are not tied to a self-consistent IMF prescription.","rationale":"The reader correctly identifies that the detectability result hinges on the Pop. III LX/SFR being elevated above the Pop. II value; the paper even demonstrates this by showing the weak model is indistinguishable. My stress-test pass agrees with that core concern but sharpens it: the high-X-ray models are not merely unconstrained, they may be inconsistent with the IMFs to which they are attached. The paper's own Appendix B notes the IMF–X-ray coupling for the log-normal IMF, yet the moderate model uses a Salpeter IMF with a 10x boost without justifying that combination. If the Sartorio et al. relation implies a Salpeter IMF produces only ~3e40, then the moderate model is an artifact of the chosen parameter grid, and the detectable claim reduces to a single compounded extreme model. This strengthens the need for the revision the reader already requested: tie LX/SFR to the IMF or clearly label the grid as a toy exploration. However, the paper is transparent about the conditional nature, states the weak model is undetectable, and provides the model comparisons needed to interpret the result. Therefore the reader's CONDITIONAL verdict remains appropriate; no change is required. The proposed test would settle whether the moderate model is physically grounded and, if not, whether the detectable regime is narrower than presented.","tokens_in":22015,"tokens_out":12488,"duration_ms":133697,"concrete_test":"Run the Sartorio et al. (2023) population-synthesis pipeline to compute the self-consistent L_X<2keV/SFR for the Salpeter and log-normal IMFs adopted in Table 4, using the same metallicity, binary fraction, and spectral models as that work. Then re-run the moderate and extreme models with these self-consistent X-ray luminosities and recompute D (Eq. 6) for z = 7, 8, 9, 10 and the same 1080 h SKA1-low campaign. If no model with a self-consistent LX/SFR exceeds the Pop. II value by more than ~1 dex, or if the moderate model's Salpeter IMF yields a value below 3e41, the claimed SKA distinguishability ceases to correspond to any physically motivated model.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim — that models with Pop. III X-ray emission stronger than Pop. II are distinguishable with ~1000 h of SKA1-low — is conditional on the specific values of L_X<2keV,III/SFR used in Table 4: 3e41 and 3e42 erg s^-1 M_sun^-1 yr for the moderate and extreme models, respectively. The paper itself shows that the weak model (3e40, comparable to the fiducial Pop. II value) yields D<1 at all z and k even with 1080 h, so no detectability. The moderate and extreme values are motivated by Sartorio et al. (2023), but the paper treats LX/SFR as an independent parameter rather than deriving it from the adopted IMF and stellar-physics assumptions. In particular, the moderate model pairs a Salpeter IMF with a 10x Pop. II X-ray luminosity, while the Appendix B discussion acknowledges that for a top-heavy IMF, linking LX/SFR to IMF would require an enhanced X-ray emissivity to be realistic. This raises the question of whether any self-consistent Pop. III model actually lands in the detectable regime, or whether the physical LX/SFR range is narrower than the explored grid. If the self-consistent value for a Salpeter IMF is close to the Pop. II value, the only remaining detectable case is the extreme model, which compounds three parameter changes and is therefore less robust as a clean attribution. Consequently, the paper's observational forecast is load-bearing on an untested and possibly internally inconsistent mapping between IMF and X-ray output.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper extends the meraxes semi-analytic model of Pop. III star formation to a large 210 h^-1 cMpc cosmological volume by calibrating density-dependent scaling relations for the mini-halo star formation rate on a smaller, higher-resolution 10 h^-1 cMpc simulation. The authors compute the 21cm global signal and power spectrum for four Pop. III models (weak, moderate, extreme, and a high-SFE variant) and use 21cmSense to forecast the detectability of Pop. III X-ray heating with SKA1-low. Their central finding is that Pop. III stars have little effect on reionization but can significantly heat the IGM at z >= 15, boosting the 21cm power spectrum at z <= 10; models with Pop. III X-ray emission stronger than Pop. II are distinguishable from models with no or mild Pop. III emission with about 1000 hours of SKA1-low observations.","tokens_in":22416,"tokens_out":7169,"duration_ms":65439,"significance":"The methodological advance is the scaling-relation approach that extrapolates mini-halo star formation from a high-resolution small box to a reionization-sized volume. The method is validated with K-S tests and reproduces the L10 mini-halo SFRD to within about 10% at all redshifts. The self-consistent coupling of Pop. III star formation, reionization, and 21cm physics in a large volume is a useful step beyond earlier analytic treatments. The paper is also careful to avoid circularity: the model is calibrated to non-21cm observables, so the 21cm forecasts are genuine predictions rather than fits. However, the headline detectability claim is conditional on the assumed specific X-ray luminosity of Pop. III stars, a parameter that is unconstrained by observation and not yet tied to the adopted IMF in a self-consistent way.","major_comments":[{"comment":"The central SKA detectability result rests entirely on the assumed specific X-ray luminosity values L_X<2keV,III/SFR = 3e41 and 3e42 erg s^-1 M_sun^-1 yr for the moderate and extreme models. The paper's own weak model (3e40, comparable to the Pop. II fiducial) yields D < 1 at all z and k even with 1080 hours (Fig. 14), so the forecast is purely conditional on these unconstrained inputs. The moderate model pairs a Salpeter IMF with a 10x enhanced X-ray luminosity, while the extreme model compounds a log-normal IMF, a tenfold SFE increase, and a 100x X-ray luminosity; Appendix B explicitly concedes that the LogE model is not realistic when L_X/SFR is linked to the IMF. Since Sartorio et al. (2023) predict the X-ray emissivity to be IMF-dependent, the paper needs either to compute L_X/SFR self-consistently from stellar-population synthesis for each IMF, or to reframe the SKA claim as an upper-limit forecast under optimistic assumptions rather than a prediction for plausible Pop. III models.","section":"Table 4, Sec. 5, Eq. (6)"},{"comment":"The scaling-relation reconstruction is validated only against the angle-averaged SFRD (ratio within 10%). The 21cm power spectrum is a two-point statistic that depends on the spatial distribution of the reconstructed SFR field, not just its mean. The L210 density field is computed from particles of mass 3.16e7 M_sun, while the L10 uses 4.71e5 M_sun; if the small-scale density statistics differ, the reconstructed SFR field may have incorrect clustering even when the mean SFRD matches. I request a validation of the reconstructed SFR field power spectrum (or the corresponding 21cm signal) against the direct L10 simulation output, or an explicit demonstration that any spatial mismatch is subdominant to the differences between the Pop. III models considered.","section":"Section 3.1, Figs. 5-6"}],"minor_comments":[{"comment":"The last column header reads 'L_X<2keV,II/SFR' but the entries are Pop. III values; rename to 'L_X<2keV,III/SFR'.","section":"Table 4"},{"comment":"The abstract states 1000 hours of SKA1-low while the text and figure captions quote 1080 hours; please make the observing times consistent.","section":"Abstract and Sec. 5"},{"comment":"The sentence 'if a pixel is irradiated by a LW flux J_LW >= J_crit above a critical threshold defined as M_crit,MC = M_ato' is garbled; please rephrase to clarify that M_crit,MC is set to the atomic cooling mass when J_LW exceeds the threshold.","section":"Section 3.1, footnote 2"},{"comment":"The p-value histogram would be more interpretable if the text stated the number of cells entering the K-S test for each overdensity bin or the total number of tests performed.","section":"Figure 2"}],"recommendation":"major_revision","confidential_remarks":null},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The genuinely new thing here is the scaling-relation method: populating mini-halo star formation in a 210 Mpc/h box from a 10 Mpc/h calibration box, then computing the 21cm signal self-consistently. That is a real technical contribution, and it is validated about as well as it can be without a direct simulation at that volume—the K-S tests and the ~10% agreement in reconstructed SFRD give me confidence the method is sound. The paper is also honest in its framing: the 21cm power spectra are not used to calibrate anything, so the result is conditional rather than circular, and the authors explicitly flag that the weak Pop. III model is indistinguishable from no Pop. III even with 1080 hours of SKA1-low.\n\nThe soft spot is the one the stress-test note lands on, though it is a bit less damning than it first appears. The detectability claim for the moderate and extreme models rests entirely on LX/SFR values of 3e41 and 3e42 erg/s per solar mass per year, taken from Sartorio et al. (2023). These are not derived from the IMF in a self-consistent way, and the authors themselves admit in Appendix B that the LogE model with low LX/SFR is unrealistic if the X-ray emissivity is tied to the IMF. So they are aware of the tension. Still, the forecast is load-bearing on an assumed parameter range rather than a physically motivated one. If Pop. III X-ray emission is comparable to Pop. II, the SKA detection fails. That is a real limitation, but not a fatal one from a modelling perspective: the paper is clear that the scenario is conditional, and the qualitative physics—early X-ray heating leaves a residual imprint at z<=10—is robust.\n\nTwo other quibbles: the extreme model varies SFE, IMF, and LX/SFR simultaneously, so it cannot cleanly attribute the effect to any one parameter, and the power spectra do not include propagated model uncertainties, only observational noise. These are minor-to-moderate and fixable in revision.\n\nWho is this for? Anyone working on 21cm cosmology, EoR modelling, or Pop. III constraints. The scaling-relation technique will be used by other groups, and the forecast sharpens an SKA target. It deserves a serious referee, and I would accept it with a request to either derive LX/SFR from the IMF or clearly frame the moderate and extreme models as exploratory bounds rather than predictions.","headline":"Solid, transparent modelling study with a genuine methodological step (scaling relations for mini-halo SFR), but the SKA detectability claim hinges on unconstrained LX/SFR values; worth refereeing with a request to sharpen the parameter justification.","tokens_in":22958,"tokens_out":1484,"would_cite":true,"duration_ms":16910,"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":"Population III stars barely affect reionization but heat the early intergalactic medium through X-rays, and a 1000-hour SKA1-low campaign can see the imprint in the 21cm power spectrum at redshift 7 to 10.","keywords":["21cm cosmology","Population III stars","reionization","X-ray heating","intergalactic medium","SKA1-low","semi-analytic models","cosmic dawn"],"falsifier":"A 21cm power spectrum measurement with SKA1-low at $z = 7$ to $10$ that matches the no-Pop.-III model within the forecast errors, or any independent constraint placing the Pop. III soft X-ray luminosity per unit star formation at or below $3 \\times 10^{40}$ erg s$^{-1}$ per solar mass per year, would falsify the central detectability claim.","tokens_in":21792,"feed_emoji":"📡","tokens_out":8271,"duration_ms":68473,"temperature":0.7,"pith_summary":"Population III (Pop. III) stars, the first metal-free stars, are thought to form in mini-halos too small to resolve in large cosmological simulations, leaving their influence on the 21cm signal unknown. This paper develops a semi-analytic framework that places Pop. III star formation into a $210\\,h^{-1}\\,\\mathrm{cMpc}$ simulation via scaling relations calibrated on a high-resolution box, and computes the resulting 21cm global signal and power spectrum. The central finding is that Pop. III stars barely change the reionization history but, through X-rays from their remnants, heat the neutral intergalactic medium at $z \\geq 15$, boosting the 21cm power spectrum at $z \\leq 10$. The authors forecast that with 1000 hours of SKA1-low observations, models with Pop. III X-ray emission stronger than Pop. II can be distinguished from models with no or mild Pop. III emission. If correct, the 21cm power spectrum becomes a direct probe of the properties of the first stars.","feed_headline":"X-ray glow of first stars is detectable in 21cm maps","feed_subtitle":"With 1000 hours of SKA1-low, models with strong Pop. III X-ray emission separate from no-Pop.-III at redshift 7 to 10.","key_machinery":"The central machinery is the semi-analytic galaxy formation model meraxes, extended with a stochastic scaling relation that assigns Pop. III (and subsequent Pop. II) star formation rates to dark-matter overdensity pixels, calibrated on a small ($10\\,h^{-1}\\,\\mathrm{cMpc}$) high-resolution simulation that resolves mini-halos and validated to reproduce the mini-halo SFRD in the large box to within about 10 per cent. This lets the authors compute the UV, X-ray, Lyman-$\\alpha$ and Lyman-Werner backgrounds from unresolved mini-halos self-consistently with reionization, using a modified 21cmFast to get the ionization and spin-temperature fields that enter the 21cm brightness temperature. The parameter that carries the signal is the specific soft X-ray luminosity $L_{X}^{<2\\,\\mathrm{keV}}/\\mathrm{SFR}$ of Pop. III stars, which sets the early heating and hence the amplitude of the later power spectrum.","core_discovery":"On the paper's own terms, the discovery is that the residual signature of early Pop. III X-ray heating survives into the Epoch of Reionization and is visible in the 21cm power spectrum at $z \\leq 10$, even though reionization itself is unchanged. Using the semi-analytic model meraxes with a new density-field scaling relation for mini-halo star formation, the authors find that in all their models Pop. III stars contribute little to reionization: the Thomson optical depth and neutral fraction histories stay consistent with observations. Instead, the soft X-ray luminosity per unit star formation from Pop. III remnants, taken to be one to two orders of magnitude above the Pop. II value, heats the IGM at $z \\geq 15$, turning the 21cm sky-averaged signal into emission earlier ($z \\sim 13$ for moderate, $z \\sim 18$ for extreme models) and raising the power spectrum by more than a factor of four at $z \\geq 7$ on both large and small scales. Forecasts with 21cmsense for SKA1-low show that the moderate and extreme Pop. III models differ from the no-Pop. III model by several $\\sigma$ at $z = 7$--$10$ with 1000 hours, while the weak model with Pop. II-level X-ray efficiency is indistinguishable.","pith_inferences":["If a detection of enhanced 21cm power at $z \\sim 8$ is confirmed, it would indirectly constrain the Pop. III initial mass function, since the theoretically expected X-ray yield depends strongly on IMF shape and this paper treats $L_X/\\mathrm{SFR}$ independently of the IMF.","The density-field scaling relation technique could be reused to include other unresolved sub-grid processes (e.g. faint atomic-cooling halos or streaming-velocity effects) in large-box reionization simulations without re-running high-resolution hydrodynamics.","A non-detection with deep SKA observations would not rule out Pop. III star formation itself; it would only push the Pop. III X-ray efficiency toward the Pop. II value, leaving room for Pop. III stars with lower remnant X-ray output.","The same machinery could be combined with global-signal experiments to break degeneracies between $L_X/\\mathrm{SFR}$ and star formation efficiency, which the power spectrum alone leaves partially unresolved."],"forward_implications":["Pop. III star formation leaves reionization histories consistent with current Thomson optical depth and neutral fraction constraints, so existing EoR measurements do not rule out a substantial Pop. III population.","The 21cm global signal is driven into emission earlier when Pop. III X-ray heating is strong, making the absorption trough shallower and shifted to higher redshift.","The 21cm power spectrum at $z = 7$--$10$ is enhanced by more than a factor of four for strong Pop. III X-ray models, on both large and small scales.","A 1000-hour SKA1-low campaign can distinguish moderate and extreme Pop. III X-ray models from no- or mild-Pop. III models at $z = 7$--$10$, while 180 hours can already exclude the strongest model.","At $z \\leq 6$ the power spectrum is insensitive to the Pop. III models, since the ionization state dominates and is nearly identical across the models."],"supporting_citations":[{"why":"Supplies the Pop. III star formation and mini-halo model in meraxes, including IMF prescriptions and feedback, that this work extends to a large box.","marker":"Ventura et al. 2024"},{"why":"Provides the meraxes calibration for Pop. II on the 210 h^-1 Mpc box and the radiative background implementation (UV, Lyman-alpha, X-ray) used here.","marker":"Balu et al. 2023a"},{"why":"Provides the theoretical estimates of Pop. III LX/SFR that set the moderate and extreme model values.","marker":"Sartorio et al. 2023"},{"why":"Baseline comparison for 21cm power spectrum models with varying Pop. III X-ray properties.","marker":"Muñoz et al. 2022"},{"why":"Contemporaneous independent model showing similar effect of Pop. III X-ray heating on the 21cm signal.","marker":"Gessey-Jones et al. 2025"},{"why":"Provides the H2 self-shielding fitting function that updates the critical mini-halo mass in Appendix A.","marker":"Kulkarni et al. 2021"},{"why":"21cmFast, the semi-numerical code whose modified version computes reionization and thermal evolution in meraxes.","marker":"Mesinger et al. 2011"},{"why":"Current upper limits on the 21cm power spectrum that all models must satisfy and that motivate the redshift range.","marker":"HERA Collaboration et al. 2023"}],"fun_headline_variants":["First stars' X-ray afterglow imprints 21cm sky","SKA1-low can spot first stars' X-ray heating","Pop III X-rays brighten 21cm power spectrum","21cm maps reveal first stars' X-ray legacy"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The detectability claim depends on Population III remnants emitting soft X-rays at least ten times more efficiently per unit star formation than Population II stars, a ratio taken from theoretical models with no direct observational confirmation; if the true ratio is near the Pop. II value, the SKA forecast collapses even though the heating physics remains qualitatively correct.","fun_headline_variants_meta":{"raw":{"variants":["First stars' X-ray afterglow imprints 21cm sky","SKA1-low can spot first stars' X-ray heating","Pop III X-rays brighten 21cm power spectrum","21cm maps reveal first stars' X-ray legacy"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001084,"raw_usage":{"total_tokens":4625,"prompt_tokens":1132,"completion_tokens":3493,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":748,"completion_tokens_details":{"reasoning_tokens":3423}},"tokens_in":748,"tokens_out":3493,"duration_ms":27112,"temperature":1.0,"reasoning_tokens":3423,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T23:03:37.352124+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A 21cm power spectrum measurement with SKA1-low at $z = 7$ to $10$ that matches the no-Pop.-III model within the forecast errors, or any independent constraint placing the Pop. III soft X-ray luminosity per unit star formation at or below $3 \\times 10^{40}$ erg s$^{-1}$ per solar mass per year, would falsify the central detectability claim.","supporting_citations":[{"cited_title":"The mass distribution of the first stars can be determined via the 21-cm signal","cited_arxiv_id":"2502.18098","evidence_quote":"Contemporaneous independent model showing similar effect of Pop. III X-ray heating on the 21cm signal."}],"review_version":1}