{"id":"25c362c3-88cc-4de7-94d2-92565d322cc6","arxiv_id":"2505.04703","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":6,"one_line_summary":"Ultra-relativistic freeze-out during reheating produces the observed dark matter abundance across a far wider mass and coupling range than standard WIMP or freeze-in scenarios, with the light-mass region contingent on the cooling history.","lead":"A new study shows that dark matter freezing out while still relativistic during the post-inflation reheating epoch can match the observed cosmic abundance for masses spanning from sub-electronvolt up to a million GeV, while still ending up cold enough for structure formation. The mechanism fills the gap between the WIMP and freeze-in paradigms and substantially expands the searchable parameter space for dark matter experiments.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The coldness calculation in Sec. IV C assumes chemical and kinetic freeze-out coincide; elastic scattering with the bath persists to T_kd ≈ TFO/160 for k=2, n=2, leaving the DM warmer and shifting the Lyman-alpha bound upward, so the quoted sub-eV cold-DM example is not supported as stated.","rationale":"The central claim has two pillars: the relic abundance calculation and the coldness at structure formation. The relic-density derivation in Secs. IV A and IV B is internally consistent, with analytic expressions tested against the shown numerical solutions, so I do not see a comparably serious flaw there. The coldness pillar, however, rests on Eq. (77), which assumes the DM momentum distribution free-streams from the chemical freeze-out temperature TFO onward. That assumption fails in the featured heavy-mediator regime because elastic scattering with the bath continues after chemical freeze-out: at TFO the elastic rate exceeds H by a factor of order g_R/g_χ, and for k=2, n=2 it stays above H until T_kd ≈ TFO/160. During reheating the bath temperature falls more slowly than a^-1, so the DM that remains kinetically coupled until T_kd is substantially warmer at BBN and at matter-radiation equality than the free-streaming estimate. This directly affects the paper's most striking advertised result, namely that UFO during reheating can make sub-eV dark matter cold: for the explicit TFO = 10^5 GeV, TRH = 100 GeV benchmark, the Lyman-alpha bound moves from about 50 eV to about 0.2 keV. The abstract's 10^-7 GeV lower boundary therefore cannot be taken at face value without a kinetic-decoupling analysis. I still agree with the reader's CONDITIONAL verdict rather than a rejection: the mechanism and the higher-mass parameter window survive, and the issue is a missing physical step that can in principle be repaired. A full treatment would evolve the DM momentum distribution with the elastic scattering term and repeat the structure-formation and Neff constraints.","tokens_in":27240,"tokens_out":17039,"duration_ms":176580,"concrete_test":"Compute the kinetic decoupling temperature by equating n_R(T_kd)⟨σv⟩(T_kd) = H(T_kd) with n_R = g_R ζ(3)T^3/π^2, using the same ⟨σv⟩ = T^n/Λ^(n+2) and the reheating-era H(T) of Eq. (28), for the benchmark k=2, n=2, TFO = 10^5 GeV, TRH = 100 GeV. Then insert T_kd in place of TFO in the general Lyman-alpha bound, Eq. (88). If the bound becomes m_χ ≳ 0.2 keV instead of ~50 eV, the Sec. IV C 2 sub-eV claim fails and the low-mass boundary in Figs. 8-9 must be revised upward; repeating this over the full (TFO, TRH) grid would show how much of the 10^-7 GeV region survives.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's coldness calculation in Sec. IV C (Eqs. 77-88) identifies chemical freeze-out with kinetic decoupling. In the relativistic regime, chemical equilibrium is set by n_χ⟨σv⟩ = (3k/(k+2))H, but the rate for elastic scattering with the bath is n_R⟨σv⟩. At TFO, n_R ≈ (g_R/g_χ)n_χ, so the elastic rate exceeds H by roughly (g_R/g_χ)(3k/(k+2)), which is about 160 for g_R = 106.75, g_χ = 1, and k = 2. During reheating for k=2, H ∝ T^4 while n_R⟨σv⟩ ∝ T^5, so kinetic decoupling occurs at T_kd ≈ TFO/160 rather than at TFO. Between TFO and T_kd the DM momentum distribution tracks the bath temperature, which cools more slowly than a^-1; the momentum at structure formation is therefore larger by (TFO/T_kd)^{5/3} for k=2. For the quoted benchmark TFO = 10^5 GeV and TRH = 100 GeV, the Lyman-alpha bound in Eq. (87) rises from about 50 eV to roughly 0.2 keV, so the statement that 'even sub-eV DM particles would be cold' is not supported, and the 10^-7 GeV lower boundary of the relic-density plots is at least partially compromised. The higher-mass UFO windows above roughly keV survive, so the central mechanism remains viable, but the advertised light-cold-DM claim needs revision.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper systematically studies ultra-relativistic freeze-out (UFO) of dark matter during the inflaton-dominated reheating epoch. It derives the conditions for UFO, the freeze-out temperature (Eqs. 35, 37), constraints on the reheating temperature, and the resulting relic abundance from the Boltzmann equation with post-freeze-out production (Secs. III and IV). The authors distinguish UV-dominated from IR-dominated out-of-equilibrium production and show that the continued production of radiation during reheating dilutes the dark matter relative to the bath, relaxing the classic eV-mass bound. They quote a broad allowed parameter region: m_chi from about 10^-7 GeV to 10^6 GeV, T_RH from 10^-2 to 10^15 GeV, and Lambda from 10^3 to 10^14 GeV. They also analyze the dark-matter temperature at structure formation and the Neff constraint (Sec. IV C) to argue that the relic can be cold.","tokens_in":27605,"tokens_out":10585,"duration_ms":97414,"significance":"If the result holds, the paper opens a substantial new region of dark-matter parameter space, bridging the WIMP and FIMP paradigms. The analytic derivations are detailed and are compared with numerical Boltzmann solutions in representative cases (Figs. 4 and 7); the authors explicitly acknowledge the O(30) deviation in the m_chi > T_RH analytic estimate. The UV/IR distinction is a useful organizing principle. However, the coldness analysis in Sec. IV C relies on an implicit identification of chemical freeze-out with kinetic decoupling that is not justified and affects the advertised light-cold-DM boundary. The higher-mass UFO windows (above roughly a keV) are not affected by this flaw, so the central mechanism remains viable, but the abstract's mass range and the 'sub-eV cold DM' example require revision after a proper treatment of kinetic decoupling.","major_comments":[{"comment":"The derivation of the DM temperature at structure formation assumes that the DM free-streams from the chemical freeze-out temperature T_FO onward. This is not justified, because the elastic scattering rate of DM with the SM bath, n_R <sigma v>, exceeds the Hubble rate at T_FO by about (g_R/g_chi)(3k/(k+2)), which is roughly 160 for g_R=106.75, g_chi=1, and k=2. For k=2 during reheating H ~ T^4 while the elastic rate ~ T^5, so kinetic decoupling occurs at T_kd ~ T_FO/160 rather than at T_FO. Between T_FO and T_kd the DM momentum distribution is held at the bath temperature, which cools as a^{-3/8} (for k=2) rather than a^{-1}; at later times the momentum is larger than the free-streaming estimate by (T_FO/T_kd)^{5/3}. For the quoted benchmark T_FO=10^5 GeV and T_RH=100 GeV in Sec. IV C 2, the Lyman-alpha bound in Eq. (87) rises from about 50 eV to roughly 0.2 keV, so the statement that 'even sub-eV DM particles would be cold' is not supported as written. The abstract's lower boundary m_chi ~ 10^-7 GeV ~ 100 eV is therefore at least partially compromised. The authors should compute T_kd from n_R <sigma v> = H and evolve the DM temperature from that point, or restrict the coldness claim to masses above roughly the keV scale. The higher-mass windows are likely unaffected, but this is a load-bearing correction for the paper's light-cold-DM claim.","section":"Sec. IV C 1, Eq. (83)"},{"comment":"The same assumption of simultaneous chemical and kinetic decoupling enters the Neff bound. If kinetic decoupling occurs at T_kd < T_FO, the DM temperature at BBN is larger than the free-streamed value used in Eq. (83), because the DM tracks the bath temperature during the intervening reheating period. The resulting Delta Neff can exceed the quoted bound in parts of parameter space where T_RH is not much smaller than T_kd, so the statement that the Neff constraint is 'always satisfied' in Sec. IV C 1 needs to be re-examined after the kinetic-decoupling calculation is included. This is a direct consequence of the same premise, but it is separately relevant for the BBN-compatibility claim.","section":"Sec. IV C 1, Eq. (83)"}],"minor_comments":[{"comment":"The numerical coefficient 5 keV in Eq. (87) depends on the specific Lyman-alpha bound v_chi < 2x10^-4 at T ~ 1 eV and on the chosen degrees of freedom. The authors should specify the conversion and the assumed g_* values so the reader can track the provenance of this coefficient.","section":"Sec. IV C 2, Eq. (87)"},{"comment":"The analytic estimate for the m_chi > T_RH regime is a factor of ~30 below the numerical solution, as the authors note. Because the abstract and the parameter-space figures rely on these analytic formulas, the text should explicitly state that the analytic results in this regime are approximate at the O(30) level, even if the origin of the discrepancy is understood.","section":"Sec. IV A 3, Eq. (68) and Fig. 7"},{"comment":"The quantity n_eq^chi(a_RH) in Eq. (58) is not defined in the text; it would be clearer to write n_eq(a_RH) and define the convention (the equilibrium number density of a massless scalar with g_chi degrees of freedom).","section":"Sec. IV A 2, Eq. (58)"},{"comment":"The identification Lambda = M/g in Eq. (36) is made via Eq. (5), but the low-energy matching of Lambda to the mediator mass and coupling is stated somewhat abruptly. A one-sentence reminder of the matching in the low-temperature limit would improve readability.","section":"Sec. III B, after Eq. (35)"},{"comment":"The cases with k > 7 are repeatedly left as 'an exercise for the motivated reader.' Since the paper already derives T_FO and the T_RH constraints for k > 7, a few formulae showing the resulting relic abundance for the k=8 example would make the paper self-contained for readers interested in those potentials.","section":"Secs. IV A 3 and IV B"}],"recommendation":"major_revision","confidential_remarks":null},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Hi,\n\nWorth knowing: this paper opens a genuinely new dark-matter production window — ultra-relativistic freeze-out during inflaton-dominated reheating — and maps out a large slice of parameter space for it. The relic-density machinery in Sec. IV is internally consistent and checked against numerics in representative cases. The UV/IR UFO distinction and the observation that UFO couplings sit between the WIMP and FIMP ranges are the real contributions. The authors build directly on their own earlier reheating and freeze-in work, and that scaffolding is sound.\n\nThe soft spot is the coldness argument in Sec. IV.C. They assume the DM free-streams immediately after chemical freeze-out, so the momentum redshifts as a^{-1} from TFO. But at TFO the elastic scattering rate with the SM bath is still a factor of order 100 above Hubble for the k=2, n=2 benchmarks, so kinetic decoupling happens later, near TFO/160 rather than at TFO. During reheating the bath temperature falls slower than a^{-1}, so the DM momentum distribution stays warmer than Eq. (77) implies. The stress-test estimate moves the Lyman-alpha bound from about 50 eV to roughly 0.2 keV for their headline example. That means the sub-eV cold-DM claim is not supported as stated. The higher-mass UFO window above roughly keV survives, so the central mechanism is still viable, but the light-mass boundary needs a proper kinetic-decoupling analysis before those plots can be trusted.\n\nMinor points: the k>7 relic abundance is left to the reader, and the UFO/WIMP boundary is approximate. The authors flag both, so they are not hidden weaknesses. No code or data, but the analytic formulas are clear enough to reproduce.\n\nBottom line: this is a useful paper for DM phenomenologists and reheating cosmologists, and it deserves referee time. The authors should be asked to either justify the simultaneous chemical/kinetic decoupling assumption or redo the coldness calculation with the elastic-scattering epoch included. I would send it out, but with a clear request for revision on Sec. IV.C.","headline":"Solid new UFO mechanism, but the light-DM coldness claim assumes away the kinetic decoupling epoch — the sub-eV boundary needs a real calculation.","tokens_in":28186,"tokens_out":2225,"would_cite":true,"duration_ms":23163,"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":"Ultra-relativistic freeze-out during reheating can produce the observed dark matter abundance across thirteen orders of magnitude in mass while keeping the relic cold enough for structure formation.","keywords":["ultra-relativistic freeze-out","reheating","dark matter relic density","freeze-in","inflaton decay","cold dark matter","Lyman-alpha constraint","heavy mediator"],"falsifier":"Compute the elastic scattering rate $\\Gamma_{\\rm el} = n_R\\langle\\sigma v\\rangle$ below $T_{\\rm FO}$ for the $k=2,n=2$ benchmark and evolve the dark matter momentum distribution with kinetic decoupling treated separately; if $\\Gamma_{\\rm el}$ remains above $H$ until $T \\approx T_{\\rm FO}/100$, the Lyman-$\\alpha$ mass bound rises well above the quoted 5 keV and the sub-eV dark matter benchmark is excluded.","tokens_in":26973,"feed_emoji":"🌌","tokens_out":9582,"duration_ms":93123,"temperature":0.7,"pith_summary":"Most production mechanisms either assume dark matter is cold at decoupling (WIMP) or barely interacts (FIMP), and relativistic freeze-out is normally rejected because it gives hot eV-scale relics. The paper argues that if decoupling happens during reheating, this rejection fails: the inflaton keeps pouring radiation into the bath after dark matter freezes out, diluting the dark matter relative to the radiation and letting it cool while the bath is replenished. The result is that ultra-relativistic freeze-out during reheating can match the observed relic density over a huge parameter space, with dark matter masses from $10^{-7}$ GeV to $10^{6}$ GeV and interaction scales from $10^{3}$ to $10^{14}$ GeV, while remaining cold by structure formation. Because the couplings sit between those of WIMPs and FIMPs, the mechanism is also a natural target for detection experiments.","feed_headline":"Hot freeze-out in reheating can still make cold dark matter","feed_subtitle":"Continued inflaton decay dilutes the relic, so masses from 10^-7 to 10^6 GeV can match the observed abundance.","key_machinery":"The machinery is the Boltzmann equation for the comoving number density $Y_\\chi = n_\\chi a^3$ after decoupling, $dY_\\chi/da = a^2\\langle\\sigma v\\rangle n_{\\rm eq}^2/H(a)$, integrated from the freeze-out scale $a_{\\rm FO}$ to $a_{\\rm RH}$ during reheating. The temperature evolution is controlled by the inflaton potential exponent $k$, giving $T \\propto a^{-3(k-1)/(2(k+2))}$ and $H(T) \\propto T^{2k/(k-1)}$ for $k<7$; this makes the source term grow with $a$ even after $\\Gamma < H$, so dark matter production continues in freeze-in fashion while the bath is replenished by inflaton decay. Critical exponents $n_c$ and $n_*$ separate whether UFO is possible and whether the post-freeze-out production is UV- or IR-dominated, and the analytic relic formulas for $k=2,n=2$ and $k=4,n=2$ are the paper's concrete outputs.","core_discovery":"The paper's central claim is that dark matter which reaches thermal equilibrium during reheating and freezes out while still relativistic—ultra-relativistic freeze-out, or UFO—is a complete, self-consistent production mechanism. After freeze-out the annihilation term drops out, but the production term $n_{\\rm eq}^2\\langle\\sigma v\\rangle$ keeps feeding dark matter; simultaneously the inflaton continues to decay, increasing the comoving radiation density. The net effect is a relative dilution $n_\\chi/n_R \\propto (T_{\\rm FO}/T_{\\rm RH})^3$ for UV-dominated cases, which breaks the classical relation between dark matter mass and relic density and lets masses far above the eV scale satisfy $\\Omega_\\chi h^2=0.12$. The paper derives the conditions under which UFO is possible, the freeze-out temperature, and the resulting relic abundance, and it shows that the dark matter is cold enough at structure formation to dodge Lyman-$\\alpha$ and $N_{\\rm eff}$ constraints, distinguishing UV-UFO, whose abundance is sensitive to $T_{\\rm FO}$, from IR-UFO, whose abundance is sensitive to $m_\\chi$ and $T_{\\rm RH}$ but not to $T_{\\rm FO}$.","pith_inferences":["Beyond the paper: if kinetic decoupling is delayed below the chemical freeze-out temperature, the dark matter temperature evolution used in Sec. IV C underestimates the velocity dispersion; treating elastic scattering separately would tighten the Lyman-$\\alpha$ bounds and may remove the paper's sub-eV examples.","Beyond the paper: the same dilution logic should apply to semi-relativistic freeze-out, so the boundary between UFO and WIMP-like freeze-out is likely a smooth ramp rather than a sharp line; the paper's analytic boundaries could be tested with full numerical Boltzmann solvers.","Beyond the paper: the model-independent parametrization $\\langle\\sigma v\\rangle = T^n/\\Lambda^{n+2}$ suggests UFO should be implemented in concrete portal models and checked against future cosmological surveys of the small-scale matter power spectrum, which are more sensitive than the simple 5 keV velocity criterion.","Beyond the paper: a late enough freeze-out would leave a small thermal velocity in the dark matter today; the absence of any cutoff in the small-scale power spectrum would push $T_{\\rm FO}/T_{\\rm RH}$ to large values and favor the UV-dominated regime."],"forward_implications":["The standard ceiling $m_\\chi \\lesssim 100$ eV for relativistic freeze-out is replaced by a much wider band: with reheating-era dilution, masses from $10^{-7}$ GeV to $10^6$ GeV can give the observed relic density.","Heavy-vector or heavy-scalar mediator models with intermediate couplings will generically pass through WIMP-like freeze-out, UFO, and freeze-in as $\\Lambda$ and $T_{\\rm RH}$ vary, so UFO is not a separate model class but a contiguous regime.","UFO couplings lie between typical WIMP and FIMP values, giving the dark sector a realistic chance of being seen in direct and indirect searches, unlike FIMPs.","UV-UFO and IR-UFO relics have different fingerprints: a UV abundance depends on the freeze-out temperature, while an IR abundance depends on $m_\\chi$ and $T_{\\rm RH}$, so measuring the dark matter mass and interaction scale could identify which regime operated.","Because dark matter equilibrates before freeze-out, the final abundance does not depend on its production history during or after inflation, removing the initial-condition problem of freeze-in."],"supporting_citations":[{"why":"Derives the reheating-era scaling of $H(T)$ and $T(a)$ used to compute freeze-out and dilution.","marker":"[4]"},{"why":"Supplies the radiation-density evolution and $T_{\\rm max}$ formulas during reheating.","marker":"[5]"},{"why":"Establishes the freeze-in relic calculation during reheating that UFO is contrasted with and whose Boltzmann solution is adapted.","marker":"[6]"},{"why":"Analyzes WIMP-like non-relativistic freeze-out during reheating, the neighboring regime that marks the UFO/WIMP boundary.","marker":"[8]"},{"why":"Sets the BBN lower bound $T_{\\rm RH} \\geq 4$ MeV used to delimit allowed reheating temperatures.","marker":"[15]"},{"why":"Provides the $N_{\\rm eff}$ bound that limits extra dark radiation at BBN.","marker":"[18]"},{"why":"Fixes the observed relic density $\\Omega_\\chi h^2 = 0.12$ used throughout.","marker":"[19]"},{"why":"Supplies the Lyman-$\\alpha$ forest bound used to require cold dark matter at structure formation.","marker":"[33]"}],"fun_headline_variants":["Hot freeze-out during reheating still yields cold dark matter","Relativistic freeze-out: a broad new path to dark matter","UFO dark matter: hot freeze-out, cold today","Relativistic freeze-out allows strong DM couplings","Relativistic freeze-out spans 13 orders in DM mass"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The calculation lets dark matter free-stream from the instant of chemical freeze-out, assuming kinetic decoupling happens at the same temperature; if elastic scattering with the bath continues below $T_{\\rm FO}$, the relic is warmer and the paper's coldness and $N_{\\rm eff}$ conclusions tighten.","fun_headline_variants_meta":{"raw":{"variants":["Hot freeze-out during reheating still yields cold dark matter","Relativistic freeze-out: a broad new path to dark matter","UFO dark matter: hot freeze-out, cold today","Relativistic freeze-out allows strong DM couplings","Relativistic freeze-out spans 13 orders in DM mass"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000909,"raw_usage":{"total_tokens":4036,"prompt_tokens":1200,"completion_tokens":2836,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":816,"completion_tokens_details":{"reasoning_tokens":2754}},"tokens_in":816,"tokens_out":2836,"duration_ms":19450,"temperature":1.0,"reasoning_tokens":2754,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T23:25:18.802834+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the elastic scattering rate $\\Gamma_{\\rm el} = n_R\\langle\\sigma v\\rangle$ below $T_{\\rm FO}$ for the $k=2,n=2$ benchmark and evolve the dark matter momentum distribution with kinetic decoupling treated separately; if $\\Gamma_{\\rm el}$ remains above $H$ until $T \\approx T_{\\rm FO}/100$, the Lyman-$\\alpha$ mass bound rises well above the quoted 5 keV and the sub-eV dark matter benchmark is excluded.","supporting_citations":[{"cited_title":"However, inflaton decay continues to steadily increase the co-moving number density of SM particles during reheat- ing","cited_arxiv_id":null,"evidence_quote":"Derives the reheating-era scaling of $H(T)$ and $T(a)$ used to compute freeze-out and dilution."},{"cited_title":"radiation dominated","cited_arxiv_id":null,"evidence_quote":"Supplies the radiation-density evolution and $T_{\\rm max}$ formulas during reheating."},{"cited_title":"Indeed, when T < mχ, the exponentially suppressed Boltzmann factor significantly reduces the number of targets in the neq×neq⟨σv⟩ pro- duction term","cited_arxiv_id":null,"evidence_quote":"Establishes the freeze-in relic calculation during reheating that UFO is contrasted with and whose Boltzmann solution is adapted."},{"cited_title":"Classical","cited_arxiv_id":null,"evidence_quote":"Analyzes WIMP-like non-relativistic freeze-out during reheating, the neighboring regime that marks the UFO/WIMP boundary."}],"review_version":1}