{"id":"680a4632-eaa7-4949-9585-20122121b8bb","arxiv_id":"2411.10626","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":7,"one_line_summary":"Dark matter annihilation changes mini-halo gas content and molecular cooling, shifting the Cosmic Dawn 21cm signal in a mass- and streaming-dependent way.","lead":"This paper models how dark matter annihilation heats and ionizes gas before the first stars formed, changing how much gas small halos hold and when they can cool into stars. It predicts the 21cm radio signal shifts in timing and depth, a signature future experiments could test.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The unexplained z<25 deposition-fraction rescaling in Sec V is the most load-bearing gap: it directly enters the headline z=20 gas-fraction, Mcool, and 21cm predictions, and is neither derived nor sensitivity-tested.","rationale":"The reader's CONDITIONAL verdict is appropriate, and their list of issues already includes the fc(z<25) rescaling as item (1). I would elevate this rescaling above local deposition as the single most load-bearing concern. The local-deposition caveat is real, but it is explicitly isolated in Appendix A with a parameter h and framed as future work; the main text's central results might remain qualitatively correct for the low-mass halos where the 10-40% gas-fraction suppression is claimed. The fc rescaling, by contrast, is an unexplained modification inside the code path that produces the paper's low-redshift quantitative claims. It is exactly where a plausible bug or unjustified assumption would most directly change the abstract's numbers. A focused recomputation with and without the rescaling would settle whether it matters. The paper remains publishable conditionally, with the rescaling either derived, replaced by a validated deposition table, or demonstrated to be numerically inconsequential. No change to the reader's verdict is needed; the condition is simply sharpened.","tokens_in":30154,"tokens_out":11373,"duration_ms":114935,"concrete_test":"Rerun the pipeline for z=20-25 three ways: (i) with the paper's rescaling as specified, (ii) with the original Slatyer fc(z) tables applied directly (no rescaling), and (iii) with a validated low-z deposition calculation (e.g., DarkHistory tables). Compare fgas at Mh=1e5 Msun, Mcool/Mcool,0, and the 21cm trough redshift. If (i)-(iii) disagree by more than about 20% in fgas or Mcool, or by Delta z > 1 in the trough, the headline low-redshift predictions are not robust to this step; if they agree, the rescaling is a non-issue.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Section V modifies 21cmvFAST for z<25 with fc(z<25)=f(z')/f(z) fc(z'), where z' is the redshift at which xe(z)=xe(z'). This is the only step of the main pipeline presented without derivation, reference, or validation. It sits exactly in the epoch of the paper's headline quantitative results: the 10-40% fgas suppression at z=20 (Fig. 6), the factor-two Mcool increase at z=20 (Figs. 3/9/10), and the Delta z ~ 2-5 shift of the 21cm absorption trough (Fig. 11). The Slatyer deposition tables already encode redshift-dependent energy-deposition physics; replacing them at z<25 with an undocumented ratio evaluated at matched xe can rescale the injected heating and ionization by an uncontrolled, potentially order-unity factor. In addition, the manuscript does not state whether the z=20 semi-analytic results (Figs. 3 and 6) use the CosmoRec thermal history or the rescaled 21cmvFAST history, so the effect of this ambiguity on the headline numbers is untracked. If the rescaling is wrong, the quantitative content of the central claim changes, not just a deferred caveat.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper builds a semi-analytic model of dark matter annihilation during Cosmic Dawn (z≈20–40) and couples it to a simplified treatment of gas collapse, molecular hydrogen cooling, and star formation in mini-halos, using the public 21cmvFAST code to compute the global 21cm brightness temperature and power spectrum. The model considers s-wave annihilation of 9 MeV, 130 MeV, and 1.1 GeV dark matter into e+e− pairs, uses deposition fractions from Slatyer's tables via CosmoRec, and incorporates dark matter–baryon streaming velocities. The central results are: a 10–40% suppression of the gas fraction in halos below about 10^5 M☉ at z=20 (Fig. 6); a minimum cooling mass that decreases for z>40 but increases by up to a factor of about 2 at z=20 in the absence of streaming, with the effect reversing when streaming velocities are included (Figs. 3, 9, 10); and a shift of the global 21cm absorption trough by Δz ≈ 2–5 relative to no-annihilation models (Fig. 11).","tokens_in":30502,"tokens_out":4551,"duration_ms":45083,"significance":"If the predictions hold, the paper would provide a useful first unified estimate of DM annihilation and streaming effects on early structure formation and the 21cm signal, with a transparent analytic cooling model that reproduces the broad behavior of more expensive simulations. The use of independent energy-deposition tables, the CosmoRec thermal histories, and the public 21cmvFAST code are strengths, as is the explicit comparison with previous simulation fits for Mcool. However, the two main results—the z=20 gas fraction and the 21cm trough shift—depend on a rescaling of the deposition fraction for z<25 that is neither derived nor validated, and on the neglect of local halo energy deposition; these issues make the quantitative conclusions provisional.","major_comments":[{"comment":"The rescaling fc(z<25) = f(z')/f(z) fc(z') is introduced without derivation, reference, or sensitivity analysis. This is load-bearing because it modifies the energy deposition in exactly the epoch of the headline results (z≈20 in Figs. 6, 10, 11). The Slatyer tables already provide redshift-dependent deposition fractions; replacing them below z=25 with a ratio evaluated at a matched electron fraction can change the injected heating and ionization by an order-unity, uncontrolled factor. Please either derive this step, replace it with the unmodified tables, or show that the z<25 predictions are insensitive to it.","section":"Section V, deposition-fraction rescaling"},{"comment":"The main pipeline assumes fesc≈1, so that each halo's gas experiences only the global DM annihilation background. Appendix A itself shows that for local deposition efficiency h ≳ −64, local heating and ionization around a 10^6 M☉ halo become significant. Since the z=20 gas fraction (Fig. 6) and the sign of the DM effect on Mcool (Figs. 9–10) depend on the gas density and electron fraction in mini-halos, the neglected local channel can alter the central conclusions. Please either include local deposition in the fiducial model or state, with a quantitative estimate, the local efficiency at which the headline results change.","section":"Section IV A and Appendix A"},{"comment":"The manuscript does not state whether the z=20 semi-analytic results (gas fraction in Fig. 6 and Mcool in Figs. 3, 9, 10) use the CosmoRec thermal history or the rescaled 21cmvFAST thermal history. The text says the gas thermal history for Fig. 6 is calculated with CosmoRec, while the simulation modifies the thermal evolution with Eqs. (16)–(17) and the z<25 rescaling. This ambiguity leaves the quantitative content of the central claims untracked; please clarify which thermal history enters each figure and, if different, show the impact.","section":"Section V and Figs. 3, 6"}],"minor_comments":[{"comment":"The sentence 'The bottom panel of Fig. 1 shows the relative changes of minimum cooling mass' should refer to the bottom panel of Fig. 3.","section":"Section IV B"},{"comment":"Table I is introduced in the text but not explicitly cited; consider referencing it where the comparison with previous studies is discussed.","section":"Section II"},{"comment":"The text reads 'Galacticforeground' and should read 'Galactic foreground'.","section":"Section VII"},{"comment":"The caption states 'the H2 fractions was calculated'; the verb should agree with the plural subject.","section":"Fig. 4 caption"}],"recommendation":"major_revision","confidential_remarks":null},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Plainly: the new thing here is the combination. Individually, DM annihilation in 21cm, molecular cooling, and streaming each have a literature, but this paper is the first to put them in one semi-analytic framework and hand you Mcool(z, vbc, mDM) curves and global 21cm forecasts. The analytic cooling model in Sec IV B is transparent: the equations for xe and H2 with the annihilation ionization term are spelled out, and the baseline Mcool tracks previous fits tolerably. The two-regime behavior (annihilation raises Mcool below z~30 without streaming, but lowers it when vbc is at or above vrms) is a real, falsifiable prediction. Appendix A is an honest attempt to scope the local halo deposition that the main calculation deliberately drops; it reads as a limitation, not a dodge.\n\nSoft spots, in order. The z<25 deposition-fraction rescaling in Sec V is the one unexplained step. One sentence, no derivation, no validation, no sensitivity scan, sitting exactly on the 21cm plots where the Delta z ~ 2-5 shift claims live. It will bother a referee. Note the rescaling does not enter the analytic cooling curves or the z=20 gas fractions, which come from CosmoRec, so the paper's central claim about suppression and altered cooling does not rest on it. But the 21cm forecasts do, and the paper should say clearly which history feeds which figure and justify or drop the rescaling. Second, the neglect of local halo deposition means the Fig 6 gas-fraction suppression is a lower limit; Appendix A shows the omitted channel can reverse the story at h >= -64. That is a caveat, but it is one of the two things keeping this from being 'solved'. Third, no code or parameter files; for a paper built on public codes with a one-line modification, releasing the patch would make the forecasts auditable.\n\nCircularity is a non-issue here; nothing is fit to the 21cm signal. The citation pattern is fair, and Table I's 'what we did and did not include' is a model of honest scoping.\n\nMy take: this deserves a serious referee. The advance is modest but real, and the analytic cooling model is worth the referee's hour. Push the authors to document the rescaling, state the thermal-history provenance of each figure, and release the 21cmvFAST patch.","headline":"A genuinely new combination of DM annihilation, molecular cooling, and streaming that yields solid analytic Mcool curves, but the undocumented z<25 rescaling in the 21cmvFAST runs is the one step a referee must force out into the open.","tokens_in":31016,"tokens_out":3981,"would_cite":true,"duration_ms":41446,"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":"Dark matter annihilation at cosmic dawn can shrink the gas content of the smallest star-forming halos and shift the 21cm absorption trough, with the sign of the effect on first-star formation depending on streaming velocity.","keywords":["dark matter annihilation","cosmic dawn","21cm global signal","minihalos","molecular hydrogen cooling","dark matter-baryon streaming velocity","Population III stars","semi-analytic model"],"falsifier":"Run the same pipeline with in-halo energy deposition switched on (local deposition efficiency $h\\gtrsim -64$, the Appendix A regime) instead of $f_{\\rm esc}\\approx1$, and recompute the $z=20$ gas fractions and $M_{\\rm cool}$ curves of Figures 6 and 9; if a $10^6\\,M_\\odot$ halo then shows gas temperatures and ionizations high enough to change the sign of $\\Delta M_{\\rm cool}$, the paper's central predictions are falsified.","tokens_in":29874,"feed_emoji":"🌌","tokens_out":12866,"duration_ms":105541,"temperature":0.7,"pith_summary":"This paper argues that dark matter annihilation into electron-positron pairs during cosmic dawn ($z\\approx20$–$40$) does more than heat the intergalactic medium: it changes the supply of gas available to the first star-forming mini-halos and rewires the molecular-hydrogen cooling that lets Population III stars form. Using a semi-analytic model that feeds updated energy-deposition fractions into the CosmoRec recombination code and the 21cmvFAST simulation, the authors find that annihilation suppresses the gas fraction in halos below about $10^5\\,M_\\odot$ by 10–40% at $z=20$ and alters the minimum cooling mass $M_{\\rm cool}$ in a redshift- and mass-dependent way. The sign of the effect on star formation is not fixed: annihilation lowers $M_{\\rm cool}$ at $z>40$ but raises it by up to a factor of about 2 at $z=20$ when no streaming is present, while a dark matter–baryon velocity offset (streaming) reverses the trend and makes annihilation mostly lower $M_{\\rm cool}$. These changes propagate into a global 21cm signal whose absorption trough is shifted by $\\Delta z\\sim 2$–5 and whose power spectrum is modified, giving concrete targets for 21cm experiments. The paper thus establishes that interpreting cosmic-dawn 21cm data requires treating DM annihilation, molecular cooling, Lyman-Werner feedback, and streaming together.","feed_headline":"Cosmic dawn halos lose 10-40% of their gas to dark matter annihilation","feed_subtitle":"Annihilation heating and streaming decide when the first stars form, shifting the 21cm trough by Δz≈2-5.","key_machinery":"The load-bearing object is the minimum cooling mass $M_{\\rm cool}$, the lowest halo mass whose gas can cool via molecular hydrogen fast enough to form stars, here set by the criterion $t_{\\rm cool}<0.2\\,t_H$ with $t_{\\rm cool}$ from the cooling-time formula. Around this object the paper assembles a chain of semi-analytic pieces: the filtering mass $M_F$, the time-averaged Jeans mass that sets the gas fraction a halo retains; the H2 chemistry with DM ionization added through $\\Lambda_{\\rm ion|DM}$; the deposition fractions $f_c(z)$ for heating, ionization, and Ly-$\\alpha$ from transfer-function tables; and a boost factor $B(z)$ that accounts for annihilation power from collapsed halos. Streaming enters by replacing the effective sound speed with $c_s^{\\prime2}=c_s^2+v_{\\rm bc}^2$ and by raising the IGM temperature seen by halo gas, which lowers core gas densities. $M_{\\rm cool}$ then controls the collapsed baryon fraction passed to 21cmvFAST, so every downstream prediction—gas fractions, star-formation timing, and the 21cm signal—flows through this one threshold.","core_discovery":"The paper's central claim is that dark matter annihilation in the 10 MeV–GeV mass range, with $\\langle\\sigma v\\rangle/m_{\\rm DM}=10^{-27}\\,\\mathrm{cm^3\\,s^{-1}\\,GeV^{-1}}$ and the $\\chi\\chi\\to e^+e^-$ channel, leaves a compound imprint on the first structures. Energy deposited into the IGM raises the gas temperature and ionization fraction, which increases the time-averaged Jeans (filtering) mass $M_F$ and, through the gas-fraction prescription, reduces $f_{\\rm gas}$ in low-mass halos by 10–40% at $z=20$ for $M_h<10^5\\,M_\\odot$. In the analytic molecular-cooling model, DM-induced ionization raises the electron abundance and accelerates H2 formation at early times, while DM heating lengthens the cooling time; the competition makes $M_{\\rm cool}$ decrease slightly for $z>40$ and increase by up to a factor of about 2 at $z=20$ when streaming is absent. With a streaming velocity $v_{\\rm bc}=v_{\\rm rms}$, the gas density in mini-halos is lower, DM ionization dominates over heating, and annihilation mostly lowers $M_{\\rm cool}$ across redshifts. When these effects are propagated through 21cmvFAST, the global 21cm absorption trough is shifted by $\\Delta z\\sim 2$ from streaming and by $\\Delta z\\approx 5$ relative to earlier DM models, and the 21cm power spectrum shows altered timing and amplitude. The paper also reports that local in-halo energy deposition, if efficient, could substantially raise gas temperature and ionization around a $10^6\\,M_\\odot$ halo, a channel it leaves for future work.","pith_inferences":["We infer that if the streaming-induced sign flip is real, single-band 21cm absorption measurements cannot separate DM annihilation from streaming-induced suppression; only the joint shape of the power spectrum can break the degeneracy.","We infer that at cross-sections below the benchmark, the qualitative structure-suppression effect does not disappear but shifts to lower redshift, so a null 21cm detection would constrain the model against this redshift-dependent sensitivity.","We infer that allowing the escape fraction $f_{\\rm esc}$ to be a free parameter in the same pipeline is the natural next calculation; the Appendix A regime $h\\gtrsim -64$ suggests the predicted trough position is sensitive to local energy transport.","We infer that the fixed-temperature assumption during cooling likely overestimates the cooling rate, and a time-dependent treatment would shift $M_{\\rm cool}$ in a way testable against the simulation comparisons in the paper."],"forward_implications":["Mini-halos below about $10^5\\,M_\\odot$ at $z=20$ retain 10–40% less gas, so Population III star formation in the smallest systems is suppressed while atomic-cooling halos above $10^8\\,M_\\odot$ are essentially unaffected.","The molecular-cooling threshold is non-monotonic in redshift: DM annihilation lowers $M_{\\rm cool}$ above $z\\approx40$ and raises it by up to a factor of about 2 at $z=20$ when streaming is absent.","Turning on a streaming velocity $v_{\\rm bc}=v_{\\rm rms}$ flips the trend: annihilation mostly lowers $M_{\\rm cool}$ at most redshifts, so the same DM model that delays first stars without streaming can accelerate them with streaming.","Lyman-Werner feedback magnifies the DM effect on $M_{\\rm cool}$ at low redshift, so 21cm fits that ignore this feedback will misattribute the DM contribution.","The global 21cm absorption trough is shifted by $\\Delta z\\sim 2$ with streaming and by $\\Delta z\\approx 5$ relative to earlier no-molecular-cooling models, giving an observable discriminator."],"supporting_citations":[{"why":"Defines the dark-matter annihilation setup (electron-positron channel, MeV-GeV masses) and the 21cm baseline model this paper extends.","marker":"[17]"},{"why":"Supplies the updated energy-deposition fractions for heating, ionization, and Ly-alpha that drive the DM effects on the IGM and halos.","marker":"[18]"},{"why":"Provides the 21cmvFAST code with molecular-cooling halos and streaming-velocity treatment used for the 21cm predictions.","marker":"[21]"},{"why":"Provides CosmoRec, used for the recombination/thermal history and initial electron fractions under DM energy injection.","marker":"[55]"},{"why":"Supplies the streaming-velocity prescriptions for filtering mass, gas fraction, and cooling velocity that are incorporated and compared.","marker":"[56]"},{"why":"Supplies simulation-based critical halo mass fits for Lyman-Werner feedback and streaming used as comparison for Mcool.","marker":"[57]"},{"why":"Supplies simulation-based minimum halo mass thresholds used as comparison for Mcool.","marker":"[58]"},{"why":"Supplies the analytic H2-cooling critical mass model with a different cooling criterion, the closest comparison for the cooling model.","marker":"[54]"},{"why":"Supplies the prior dark-matter plus molecular-cooling study, compared for the 21cm signal timing.","marker":"[27]"}],"fun_headline_variants":["Dark matter annihilation slashes halo gas by 10-40% at cosmic dawn","Annihilation shifts 21cm absorption trough by up to Δz 5","DM heating and streaming control first star formation and 21cm","Cosmic dawn: DM annihilation raises Jeans mass, alters cooling","Streaming offsets amplify annihilation's impact on early halos"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The calculations assume that each halo's gas receives only the cosmic-average dark-matter annihilation background, with essentially all annihilation energy produced inside a halo escaping rather than heating that halo's own gas ($f_{\\rm esc}\\approx1$); if local deposition is efficient, the predicted gas-fraction suppression and the direction of the cooling-mass shift could change.","fun_headline_variants_meta":{"raw":{"variants":["Dark matter annihilation slashes halo gas by 10-40% at cosmic dawn","Annihilation shifts 21cm absorption trough by up to Δz 5","DM heating and streaming control first star formation and 21cm","Cosmic dawn: DM annihilation raises Jeans mass, alters cooling","Streaming offsets amplify annihilation's impact on early halos"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000256,"raw_usage":{"total_tokens":1668,"prompt_tokens":1132,"completion_tokens":536,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":748,"completion_tokens_details":{"reasoning_tokens":443}},"tokens_in":748,"tokens_out":536,"duration_ms":5887,"temperature":1.0,"reasoning_tokens":443,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T19:29:58.501807+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Run the same pipeline with in-halo energy deposition switched on (local deposition efficiency $h\\gtrsim -64$, the Appendix A regime) instead of $f_{\\rm esc}\\approx1$, and recompute the $z=20$ gas fractions and $M_{\\rm cool}$ curves of Figures 6 and 9; if a $10^6\\,M_\\odot$ halo then shows gas temperatures and ionizations high enough to change the sign of $\\Delta M_{\\rm cool}$, the paper's central predictions are falsified.","supporting_citations":[{"cited_title":"Lopez-Honorez, O","cited_arxiv_id":null,"evidence_quote":"Defines the dark-matter annihilation setup (electron-positron channel, MeV-GeV masses) and the 21cm baseline model this paper extends."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the updated energy-deposition fractions for heating, ionization, and Ly-alpha that drive the DM effects on the IGM and halos."},{"cited_title":"Glover and D","cited_arxiv_id":null,"evidence_quote":"Provides CosmoRec, used for the recombination/thermal history and initial electron fractions under DM energy injection."},{"cited_title":"Yoshida, S","cited_arxiv_id":null,"evidence_quote":"Supplies the streaming-velocity prescriptions for filtering mass, gas fraction, and cooling velocity that are incorporated and compared."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies simulation-based critical halo mass fits for Lyman-Werner feedback and streaming used as comparison for Mcool."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies simulation-based minimum halo mass thresholds used as comparison for Mcool."},{"cited_title":"Hirano, T","cited_arxiv_id":null,"evidence_quote":"Supplies the prior dark-matter plus molecular-cooling study, compared for the 21cm signal timing."}],"review_version":1}