{"id":"657be1bb-7542-4947-9305-3b9077fa06c3","arxiv_id":"2412.04385","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":5,"one_line_summary":"Using radiative-hydrodynamics simulations, this work provides the first calculation of the epoch-of-reionization thermal Sunyaev-Zel'dovich power spectrum, finding mean y of a few times 10^-8 and a small-scale signal that may dominate at l near 10^5.","lead":"Astrophysicists computed the faint Compton y-signal produced when the first stars reionized the universe, using sixteen radiation-hydrodynamics simulations. The signal is only about one percent of the galaxy cluster signal at scales probed by current experiments, but it may dominate at the arcsecond scales future CMB surveys will reach.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The ell ~ 10^5 EoR tSZ dominance prediction is not resolution-converged; it rests on rare supernova-heated cells that the paper itself identifies as resolution-sensitive, so this headline claim should be explicitly conditional on a convergence test.","rationale":"The percent-level contribution at SPT scales is well supported: the 16 simulations span different volumes, resolutions, and star-formation parameters, and the resulting mean y values (1.2-3.7 x 10^-8) bracket the analytic Hill et al. (2015) estimate, giving a crude but honest error budget. The quadratic Doppler analysis is also carefully handled, including the missing-velocity-power correction from linear theory. The weak point is the small-scale prediction. The paper's own phase-diagram analysis (Sec. 5.3, Fig. 8) ties the high-y tail to rare, very hot cells (T > 10^6-10^7 K) that are only resolved in the highest-resolution runs, and it identifies the density-temperature bifurcation in coarser runs as an unphysical resolution effect. Since the auxiliary 'high-resolution' run (10 Mpc) is not large enough to sample the same environments as CoDa II, one cannot cleanly separate resolution effects from volume or feedback effects. A controlled degradation of CoDa II to the grid resolutions of the lower-resolution runs is the cleanest test: if the high-ell excess disappears, the abstract's 'potentially dominate' claim should be downgraded; if it survives, the claim is considerably stronger. This matches the reader's conditional verdict, which I do not change.","tokens_in":26922,"tokens_out":7637,"duration_ms":71032,"concrete_test":"Degrade the CoDa II gas pressure and ionization fields used in the lightcone pipeline from their stored 2048^3 resolution to 1024^3 and 512^3, matching the lower-resolution runs in the suite, while keeping volume, cosmology, star-formation parameters, and feedback unchanged; recompute the Compton-y maps and angular power spectra. If the high-y tail (log10 y > -6) and the power excess at ell ~ 10^5 are suppressed by more than an order of magnitude compared to the 2048^3 case, the small-scale dominance claim is a resolution artifact and should be removed or heavily caveated. If the excess persists and remains above the Bolliet et al. template, the claim is supported.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The claim that the EoR tSZ power spectrum peaks and can dominate the total signal at ell ~ 10^5 (abstract; Sec. 5.1.2) is not backed by a resolution-converged result. The high-ell power comes from the high-y tail (log10 y ~ -6 to -5.5), which Sec. 5.2 attributes to gas heated by supernovae and shocks to T > 10^6-10^7 K. Sec. 5.3 and Fig. 8 show that such hot cells appear only in the higher-resolution runs (CoDa II, 10 Mpc boxes) and are absent in lower-resolution runs; the accompanying density-temperature bifurcation is explicitly labeled an unphysical resolution effect. Since the two 'high-resolution' runs also differ in volume and feedback model, the difference is not a controlled test of resolution convergence. The small-scale dominance is the paper's most novel prediction, yet it is set by the least physical, unresolved parts of the simulation (subgrid SN feedback and grid resolution), so it should be treated as conditional on a convergence test.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper presents the first systematic simulation-based calculation of the thermal Sunyaev-Zel'dovich (tSZ) signal from the Epoch of Reionization. Using sixteen RAMSES-CUDATON radiation-hydrodynamics simulations (including CoDa II and a suite of auxiliary boxes), the authors construct electron-pressure lightcones over z ≈ 6–12, integrate them to obtain Compton y maps, and compute angular power spectra. The main results are mean y values of a few × 10^-8, a tSZ contribution of roughly one percent of the cluster-level signal at SPT scales (ℓ ∼ 10^3–10^4), a quadratic Doppler contribution of order 10% of the EoR tSZ signal, and a predicted rise in the EoR tSZ power spectrum at ℓ ∼ 10^5 where it may dominate over the post-reionization signal. The paper also separates the contributions of density, temperature, and ionization fluctuations and discusses phase diagrams of the simulated gas.","tokens_in":27144,"tokens_out":13848,"duration_ms":135687,"significance":"If the results hold, this is a useful first quantification of a previously neglected EoR contribution to a standard CMB foreground, with direct relevance for current SPT-scale analyses and for future high-resolution CMB experiments. The paper's strengths include the unusually large suite of calibrated simulations, the use of independent reionization observables (neutral fraction, optical depth, photoionization rate, star formation rate) rather than the tSZ signal itself for calibration, the explicit treatment of missing large-scale velocity power in the quadratic Doppler estimate, and the honest reporting of resolution-related limitations in Section 5.3. The central sub-dominance claim at ℓ < 10^4 appears robust across the simulation suite and is consistent in order of magnitude with the analytic estimate of Hill et al. (2015).","major_comments":[{"comment":"The headline claim that the EoR tSZ signal 'peaks and can potentially dominate the total signal' at ℓ ∼ 10^5 is not supported by a resolution-converged result. The high-ℓ power is produced by the high-y tail (log10 y ∼ -6 to -5.5), which Sec. 5.2 attributes to gas heated by supernovae and shocks to T > 10^6–10^7 K. However, Sec. 5.3 and Fig. 8 explicitly describe the density-temperature bifurcation as an '(unphysical) resolution effect' that disappears as resolution increases, and note that such hot cells appear only in the higher-resolution runs. Since the two high-resolution runs (CoDa II and the 10 Mpc boxes) differ in volume, resolution, and feedback model, they do not constitute a controlled convergence test. The small-scale dominance prediction should either be backed by a fixed-physics resolution study or explicitly qualified as a simulation-dependent upper bound.","section":"Abstract; Sec. 5.1.2; Sec. 5.2; Sec. 5.3"},{"comment":"There appears to be a factor-of-two error in the conversion from gas density to electron number density. For the stated definition of x_HII as an ionized fraction in [0,1], full ionization gives x_HII = 1, but Eq. (11) then yields ρe = ρgas/2, implying n_e = ρgas/(2 m_p) for pure hydrogen. The correct electron number density for fully ionized hydrogen is n_e = ρgas/m_p; the factor 1/(1+x_HII) is appropriate for converting gas pressure to electron pressure, not for converting gas density to electron density. This same factor enters the analytic uniform-IGM estimate in Eqs. (18)–(19), so the quoted reference value y ≈ 4.22 × 10^-8 (and the helium-corrected 3.46 × 10^-8) is likely too low by a factor of about two for T = 30,000 K. The authors should clarify the definition of x_HII and correct the density-lightcone and analytic estimates, or explicitly justify why the factor 1/(1+x_HII) is present in the density conversion.","section":"Sec. 4.2, Eqs. (11)–(12); Sec. 4.3.1, Eqs. (18)–(19)"},{"comment":"The implementation of the missing large-scale velocity power correction is ambiguous and load-bearing for the reported ~10% quadratic Doppler contribution. Eq. (25) requires the average ⟨n_e v²⟩ over each snapshot, and Eq. (31) defines the missing ⟨v²⟩, but the text does not state whether ⟨v²⟩_missing is added to each cell's v² before multiplying by the local n_e, or whether it is added as a global term ⟨n_e⟩ ⟨v²⟩_missing. These two prescriptions can differ at the tens-of-percent level because n_e is strongly clustered where the velocity field is nonlinear. Please specify the exact operation used to produce the filled markers in Fig. 10 and justify it against the alternative.","section":"Sec. 4.5, Eqs. (30)–(31); Fig. 10"}],"minor_comments":[{"comment":"The caption contains a typo: 'oour simulations' should read 'our simulations'.","section":"Fig. 6 caption"},{"comment":"The template used for comparison is introduced as Shaw et al. (2010) in Sec. 4.1 but the results and figures use the Bolliet et al. (2018) template; please make the template references consistent throughout.","section":"Sec. 4.1 vs Sec. 5.1.2"},{"comment":"The CoDa II lightcones are constructed from the coarsened 2048^3 stored pressure field, not the native 4096^3 grid; this effective resolution should be stated where the 'full-resolution' maps and high-ℓ power spectra are presented, since it sets a hard limit on the small-scale signal.","section":"Sec. 3.1"},{"comment":"The sigmoid interpolation between snapshots is introduced without any sensitivity test; a brief statement on the robustness of the integrated y to the choice of β and to the snapshot spacing would strengthen the methodology.","section":"Sec. 4.1, Eq. (9)"},{"comment":"The phrase 'contributing an additional ∼10% to the tSZ signal' in the abstract is potentially confusing; the authors mean 10% of the EoR tSZ signal, not 10% of the total (cluster-dominated) tSZ signal. Please rephrase for clarity.","section":"Abstract; Sec. 5.4"}],"recommendation":"major_revision","confidential_remarks":"The paper addresses a timely and previously under-explored question, and the largest-scale conclusions are likely robust. However, the small-scale dominance claim is precisely the part that depends on unverified numerical details, and the factor-of-two issue in the density-lightcone equations must be resolved before the paper is suitable for publication. A revision that fixes the density conversion, clarifies the Doppler correction, and either adds a resolution-convergence test or explicitly demotes the ℓ ∼ 10^5 claim would make this a solid MNRAS contribution."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The headline result is worth knowing: this is the first calculation of spatially resolved tSZ maps and angular power spectra from the Epoch of Reionization using full radiation-hydrodynamics simulations. They run a suite of sixteen RAMSES-CUDATON boxes, with a sensible spread in volume, resolution, and star-formation parameters, and they calibrate against independent reionization observables (neutral fraction, optical depth, photoionization rate, SFR). The mean Compton y comes out at a few x 10^-8, in the same ballpark as the analytic Hill et al. (2015) EoR estimate, so the large-scale sub-dominance claim (about 1% of the tSZ power at SPT scales) is robust across the set. The careful treatment of the quadratic Doppler term, including the correction for missing large-scale velocity power with linear theory, is a nice piece of work.\n\nThe soft spot is exactly where the stress-test puts it: the predicted high-ell peak and possible dominance at l ~ 10^5. That signal comes from rare, very hot cells (T > 10^6-10^7 K) produced by supernovae and shocks. The paper itself shows in Sec. 5.3 that these cells appear only in the higher-resolution runs and are absent in the lower-resolution ones, and it explicitly labels the density-temperature bifurcation as an unphysical resolution effect. Because the two high-resolution runs also differ in volume and feedback, this is not a controlled convergence test. So the \"potentially dominate\" claim is a genuine prediction from the simulation, but it is not yet a robust one; it should be presented as conditional on resolution convergence. I'd add two smaller points: the power spectra have no error bars, and the data availability statement (\"reasonable requests\") is not a public release. The abstract also overstates the prior state of the art by saying the EoR contribution \"has not been previously evaluated\" when Hill et al. (2015) did give a mean-y estimate; the paper itself acknowledges that later.\n\nAll of this is fixable. The large-scale result is solid, the method is clearly described, and the authors are honest about the resolution effects—they just don't follow through to a convergence test. This paper deserves a serious referee, and I would engage with it: it gives a useful reference for tSZ foreground work and a new window on the thermal state of the IGM during reionization.\n\nRecommendation: send to peer review. Ask for a resolution study or a clear caveat on the small-scale prediction, error estimates on the power spectra, and a public data release (at least the y-maps and power spectra).","headline":"First RHD-based maps and power spectra of EoR tSZ; the small-scale dominance claim is not convergence-tested.","tokens_in":27788,"tokens_out":3063,"would_cite":true,"duration_ms":49416,"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 epoch of reionization produces a thermal SZ signal of mean Compton y ~ 3-4 x 10^-8, about one percent of cluster power at survey scales and potentially dominant near multipole 10^5.","keywords":["thermal Sunyaev-Zel'dovich effect","epoch of reionization","Compton y-parameter","radiation-hydrodynamics simulations","cosmic microwave background secondary anisotropies","intergalactic medium","quadratic Doppler distortions"],"falsifier":"Run the same reionization physics at progressively higher spatial resolution and check whether the rare cells above $10^6$ K and the $y$ power at $\\ell \\sim 10^5$ converge; if the high-temperature tail shrinks or disappears with resolution, the claimed small-scale dominance is numerical. Observationally, a future arcsecond-resolution CMB survey could search for the predicted compact peaks with $y \\sim 10^{-6}$, and their absence would rule out the smallest-scale part of the claim.","tokens_in":26711,"feed_emoji":"🌌","tokens_out":9064,"duration_ms":79539,"temperature":0.7,"pith_summary":"This paper sets out to compute the thermal Sunyaev-Zel'dovich (tSZ) signal produced during the epoch of reionization, a contribution that cluster-based templates have largely ignored. Using a suite of fully coupled radiation-hydrodynamics simulations, the authors build lightcones of electron pressure over $z \\sim 6$ to 12 and integrate them along the line of sight to obtain maps and angular power spectra of the Compton $y$-parameter. They find a mean $y$ of a few times $10^{-8}$, roughly one percent of the cluster-dominated tSZ power at arcminute-resolution survey scales, with quadratic Doppler distortions adding about ten percent to that value. At much smaller scales, near multipole $\\ell \\sim 10^5$, the reionization signal peaks and can rival or exceed the post-reionization template. If the claim holds, future high-resolution CMB experiments would need to include reionization-era gas when modelling the tSZ foreground.","feed_headline":"Reionization adds a percent-level signal to the tSZ power spectrum","feed_subtitle":"New simulations find a mean Compton y of a few x 10^-8, with the reionization signal peaking at arcsecond scales.","key_machinery":"The carrying object is the Compton $y$-parameter, $y = (\\sigma_T/m_e c^2) \\int p_e\\, dl$, the line-of-sight integral of electron pressure, evaluated on lightcones built by interpolating gas pressure and ionization fraction between simulation snapshots. For a hydrogen-only gas the electron pressure is $p_e = x_{\\mathrm{HII}}/(1+x_{\\mathrm{HII}})\\, p_{\\mathrm{gas}}$, with a post-processing correction for singly ionized helium. A second-order Doppler term, $y = (\\sigma_T/3cH_0) \\int n_e\\langle v^2\\rangle\\,(1+z)^{-1}E(z)^{-1}\\, dz$, is added and corrected for velocity power missing from finite simulation boxes. These lightcone integrals convert the patchy structure of reionization directly into $y$-maps and angular power spectra.","core_discovery":"The central claim is that the epoch of reionization produces a measurable tSZ signal with mean Compton parameter $y \\sim 3$--$4 \\times 10^{-8}$, whose angular power spectrum is flatter than the cluster template and crosses it near $\\ell \\sim 10^5$. The signal grows with the redshift at which reionization completes: simulations that end reionization near $z \\sim 6.6$ reach mean $y$ values up to two or three times those of runs that finish near $z \\sim 5.6$. The high-pressure regions that drive the small-scale power are rare cells heated to millions of kelvin by supernova explosions and structure-formation shocks; lower-resolution runs smear these cells out and lose the high-$y$ tail. Isolating the density, temperature, and ionization contributions shows that the patchiness of reionization broadens the $y$ distribution, while the supernova-heated temperature fluctuations are responsible for its high-$y$ tail.","pith_inferences":["If the small-scale dominance survives higher-resolution tests, the height and shape of the EoR tSZ peak near $\\ell \\sim 10^5$ could be used to constrain supernova feedback and the clumpiness of the ionized intergalactic medium, since the peak is controlled by rare hot cells rather than by the mean IGM temperature.","The same lightcone machinery could be cross-correlated with 21-cm emission or the kinetic SZ effect; the patchy ionization and hot-cell structure should produce a distinctive cross-spectrum that separates the reionization contribution from low-redshift clusters.","Because the analytic uniform-IGM estimate ($y \\approx 4 \\times 10^{-8}$) lies close to the simulated means, a moderately improved analytic model that treats temperature fluctuations and patchiness could predict the EoR tSZ without expensive radiative-transfer simulations.","A direct extension would replace the single-ionization helium correction with a full helium reionization history; helium reionization ends later than hydrogen and could change the electron pressure at the lower-redshift end of the lightcones by an amount the current correction does not capture."],"forward_implications":["At angular scales probed by arcminute-resolution experiments ($\\ell \\sim 10^3$--$10^4$), the reionization tSZ contributes roughly one percent of the cluster-template power and should be included in foreground models for precision cosmology.","The quadratic Doppler effect adds about ten percent to the reionization tSZ $y$-parameter, with the exact fraction set by the reionization history and by large-scale velocity power missing from the simulation boxes.","Earlier reionization produces a stronger tSZ signal: mean $y$ increases roughly with the redshift at which reionization completes, with additional scatter from star formation efficiency and supernova heating.","At multipoles near $\\ell \\sim 10^5$, the reionization tSZ can dominate the post-reionization template, so future arcsecond-resolution CMB experiments would see reionization-era gas as a foreground that must be modelled.","The tSZ signal from reionization offers a redshift-independent probe of the thermal state, clumpiness, and timing of reionization, complementary to 21-cm and kinetic SZ probes."],"supporting_citations":[{"why":"Supplies the fiducial high-resolution radiation-hydrodynamics simulation whose pressure snapshots are used for the main lightcone.","marker":"Ocvirk et al. 2018"},{"why":"Provides the adaptive-mesh hydrodynamics code that the simulation suite is built on.","marker":"Teyssier 2002"},{"why":"Adds the radiative-transfer coupling used to evolve the ionized fraction and gas temperature during reionization.","marker":"Aubert & Teyssier 2008"},{"why":"Gives the previous analytic estimate of the mean EoR Compton y (9.8 x 10^-8) that the simulations are compared with.","marker":"Hill et al. 2015"},{"why":"Supplies the post-reionization tSZ power spectrum template used to judge the relative size of the EoR contribution.","marker":"Bolliet et al. 2018"},{"why":"Provides the observed CMB power spectra at 95 and 150 GHz against which the predicted signals are displayed.","marker":"George et al. 2015"},{"why":"Establishes the quadratic Doppler term that the paper evaluates as an additional ~10% contribution.","marker":"Hu et al. 1994"},{"why":"Provides an earlier tSZ template and scaling used for comparison in the power spectrum analysis.","marker":"Shaw et al. 2010"}],"fun_headline_variants":["Reionization leaves a measurable tSZ imprint on CMB maps","Epoch of reionization adds percent-level tSZ power","New tSZ simulations show reionization signal at arcsecond scales","Reionization's pressure signal could contaminate future surveys"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The predicted small-scale dominance rests on the simulation resolving rare cells heated above a million kelvin by supernova feedback and shocks; if those hot cells are numerical artefacts of resolution or if feedback is weaker, the small-scale peak would not surpass the cluster template.","fun_headline_variants_meta":{"raw":{"variants":["Reionization leaves a measurable tSZ imprint on CMB maps","Epoch of reionization adds percent-level tSZ power","New tSZ simulations show reionization signal at arcsecond scales","Reionization's pressure signal could contaminate future surveys"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000725,"raw_usage":{"total_tokens":3322,"prompt_tokens":1090,"completion_tokens":2232,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":706,"completion_tokens_details":{"reasoning_tokens":2159}},"tokens_in":706,"tokens_out":2232,"duration_ms":14678,"temperature":1.0,"reasoning_tokens":2159,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T21:24:20.913203+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Run the same reionization physics at progressively higher spatial resolution and check whether the rare cells above $10^6$ K and the $y$ power at $\\ell \\sim 10^5$ converge; if the high-temperature tail shrinks or disappears with resolution, the claimed small-scale dominance is numerical. Observationally, a future arcsecond-resolution CMB survey could search for the predicted compact peaks with $y \\sim 10^{-6}$, and their absence would rule out the smallest-scale part of the claim.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the fiducial high-resolution radiation-hydrodynamics simulation whose pressure snapshots are used for the main lightcone."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Adds the radiative-transfer coupling used to evolve the ionized fraction and gas temperature during reionization."}],"review_version":1}