{"id":"9f3ea59f-87ea-4a26-b537-fdc85bd5b637","arxiv_id":"1908.11378","paper_version":2,"verdict":"ACCEPT","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":7,"one_line_summary":"In the curvaton scenario, heavy particles with masses near the inflationary Hubble scale generate observable non-Gaussian signals because the curvaton sector can have a much lower EFT cutoff than the inflaton sector.","lead":"This paper shows that heavy particles coupled to the curvaton field, rather than the inflaton, can produce primordial non-Gaussianities orders of magnitude larger than in standard inflation while keeping the effective field theory under control. This could bring cosmological collider signatures of new particles within reach of future large-scale structure and 21-cm observations.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Fermion-loop observability depends on the sub-control benchmark Λσ=4H; with the paper's own Vσ^{1/4} bound the fermion signal drops below quoted 21-cm sensitivity.","rationale":"The paper is a self-contained and largely consistent calculation; the central mechanism—separate EFT cutoffs for the curvaton and inflaton sectors—is clearly stated and explicitly supported by proof-of-principle UV completions. The reader's weakest-assumption identification (the sequestered two-sector construction) is correct, but the most load-bearing quantitative point is more specific: the benchmark Λσ=4H used for the headline loop-level amplitudes violates the paper's own control bound (5.3), and the mediator matching that produces it is at the edge of perturbative control. The scalar loop and tree-level signals remain observable even under a more conservative cutoff, so the paper's central claim of orders-of-magnitude enhancement is not refuted. However, the fermion-loop claim, which the reader lists among the strongest results, is sensitive to this benchmark: imposing Λσ≳Vσ^{1/4}∼7H reduces |fΨ,loop| below the quoted 21-cm sensitivity. A targeted check of this cutoff dependence would settle whether the fermion-loop observability claim should be retained or softened. Because the paper's own text flags the benchmark choices and the UV completion as proof-of-principle, this concern does not warrant rejection, but it does motivate conditioning acceptance on the EFT-control check described above.","tokens_in":21560,"tokens_out":33852,"duration_ms":347610,"concrete_test":"Recompute the fermion and scalar loop amplitudes of Figs. 5-6 with Λσ=7H and Λσ=10H, keeping all other benchmark parameters fixed. In parallel, scan the mediator parameters in Eqs. (5.8)-(5.12) under the requirements MΣ>max(mχ,H) and μΣ<MΣ (or at least μΣ²/(16π²MΣ²)<0.1), and record the minimum attainable Λσ,e. If |fΨ,loop| falls below 10⁻⁴ at 21-cm sensitivity, the claim that loop-level fermion NG is observable should be downgraded, while scalar/tree-level observability can stand.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The paper's strongest quantitative claim—that loop-level NG is observable—is carried by the benchmark Λσ=4H, ˙σ0=-H², giving |fχ,loop|∼10⁻²-10⁻¹ and |fΨ,loop|∼10⁻³ (Figs. 5, 6). This Λσ is obtained in Sec. 5 by integrating out a mediator with MΣ=3H, μΣ=6H in Eq. (5.10), yielding Λσ,e²=MΣ²Λσ/μΣ≈15H², i.e. Λσ,e≈3.9H. That effective scale sits below the paper's own control condition (5.3), Λσ≳Vσ^{1/4}∼7H for the benchmark ησ=10⁻³ (Vσ≈2500H⁴). The full EFT cutoff Λσ=10H satisfies (5.3), but the low-energy theory obtained by integrating out the mediator has cutoff ∼MΣ=3H, below Vσ^{1/4}; moreover, the heavy-particle masses plotted extend to 7H, exceeding both MΣ and Λσ,e, so the effective-operator description (5.9) is not valid for much of the plotted range. The mediator choice is also marginal: μΣ/MΣ=2 with MΣ=3H is not a clean large-separation matching; a more conservative choice MΣ≳10H with μΣ<MΣ gives Λσ,e≳10H, not 4H. Rescaling the loop amplitudes from Λσ=4H to the control-bound value Λσ=7H lowers |fΨ,loop| by (4/7)⁶≈3×10⁻², i.e. to ∼3×10⁻⁵, below the σ_fNL∼10⁻⁴ sensitivity quoted in Sec. 2. Scalar-loop and tree-level signals survive this rescaling, so the central 'orders-of-magnitude enhancement' is partially robust, but the specific claim that loop-level fermion NG is observable is the least secure quantitative statement.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper revisits cosmological-collider signatures of heavy particles in a curvaton model in which the inflaton and the curvaton belong to two sequestered sectors with independent EFT cutoffs, Λφ and Λσ. It first shows that in standard single-field inflation, loop-level non-Gaussianity from charged scalars and fermions is unobservably small, with |fNL| of order 10^-9, due to the high cutoff Λφ required by control of the inflaton potential. It then argues that because the curvaton does not drive the background expansion, its couplings to heavy fields need only be suppressed by Λσ ≫ Vσ^{1/4} ~ 10H, which can be much smaller than Λφ, and that integrating out mediator fields can reduce the effective suppression scale to about 4H. For the benchmark Λσ=4H and ˙σ0=-H², the paper finds tree-level Higgs-exchange |fχ,tree| of order 0.1, scalar-loop |fχ,loop| of order 10^-2-10^-1, and fermion-loop |fΨ,loop| of order 10^-3 (Figs. 4-6), claiming that loop-level effects become observable with 21-cm-level sensitivity. The explicit one-loop calculations are carried out in Appendices A and B using the loop-to-tree reduction of Arkani-Hamed and Maldacena, and the paper also discusses observational constraints, a bi-axion monodromy completion for the large curvaton field range, and prospects for future work.","tokens_in":22110,"tokens_out":8271,"duration_ms":83776,"significance":"If the central claim holds, the paper provides a concrete and simple route to large cosmological-collider signals: it avoids the standard Λφ > 250H suppression by separating the EFT describing curvature fluctuations from the EFT describing the inflationary background. The parametric scalings |fχ,loop| ~ 1/Λσ^4 and |fΨ,loop| ~ 1/Λσ^6 make the enhancement mechanism transparent, and the explicit loop-to-tree reduction in Appendices A and B gives the paper a degree of rigor that is not always present in such phenomenological studies. I also credit the paper for clearly comparing its predictions with current Planck bounds and future 21-cm sensitivities, and for identifying the field-range issue and proposing a UV completion. The main quantitative claim, however, is tied to the benchmark Λσ=4H, and the stated control condition in Eq. (5.3) points to a more conservative cutoff around 7-10H; the fermion-loop signal is especially sensitive to this choice. The significance is therefore genuine but conditional on a controlled justification of the benchmark cutoff.","major_comments":[{"comment":"The numerical benchmark Λσ,e ≈ 4H is below the control lower bound Λσ ≳ Vσ^{1/4} ≈ 10H that the paper itself states in Eq. (5.3) for the benchmark ησ=10^-3. With the example values MΣ=3H and μΣ=6H, Eq. (5.10) gives Λσ,e ≈ 3.9H, and the residual theory after integrating out the mediator is not obviously an EFT whose cutoff lies above the curvaton energy scale. Because the fermion amplitude in Eq. (5.23) scales as Λσ^{-6}, replacing 4H by the compliant value 10H lowers |fΨ,loop| by a factor (4/10)^6 ≈ 4 × 10^-3, bringing the fermion-loop signal from ~10^-3 down to near or below the σ_fNL ≈ 10^-4 sensitivity quoted in Sec. 2. Please either justify the Λσ=4H benchmark against the control condition (5.3), or present the fermion-loop forecasts with a cutoff that satisfies (5.3).","section":"Sec. 5, Eqs. (5.3), (5.10), Figs. 5 and 6"},{"comment":"The effective operator in Eq. (5.9) is obtained by integrating out a mediator of mass MΣ=3H, and the resulting effective cutoff is Λσ,e ≈ 4H, yet the loop amplitudes in Figs. 5 and 6 are plotted for heavy masses mχ and mΨ extending up to approximately 7H. The dimension-six operator in Eq. (5.9) is only valid for external momenta and heavy-field masses below MΣ and Λσ,e; for masses above these scales the heavy field should be integrated out together with the mediator. This extrapolation affects exactly the high-mass end of the curves that is used to support the claim that loop-level effects are observable, so the predictions in this region need to be restricted or recomputed.","section":"Sec. 5, Eq. (5.9), Figs. 5 and 6"},{"comment":"The sequestered two-sector construction is the load-bearing premise of the paper, and its consistency with the presence of a common set of heavy fields {χ} in both L_int^φ and L_int^σ in Eq. (3.2) is not fully explained. If the heavy particles couple to both the inflaton and the curvaton, their inflaton couplings will also generate non-Gaussianity and may contaminate the curvaton signal or modify the scalar power spectrum beyond the quoted O(ρ1²/H²) estimate. The paper should either state explicitly that only the curvaton couplings are present in the benchmark scenario, or quantify the contribution of the inflaton couplings in the same setup.","section":"Sec. 3.1 and Sec. 5"}],"minor_comments":[{"comment":"After Eq. (5.10) the notation Λσ is reused for Λσ,e, which makes it easy to confuse the effective suppression scale with the bare cutoff in Eqs. (5.19)-(5.24). I recommend keeping Λσ,e explicitly throughout Section 5.","section":"Sec. 5, after Eq. (5.10)"},{"comment":"The tree-level bispectrum in Eq. (5.16) is quoted from [10] without a derivation or a statement of normalization relative to Eq. (2.3). Since this formula drives the tree-level plot in Fig. 4, a brief derivation or at least an explicit matching of the conventions would be helpful.","section":"Sec. 5.1, Eq. (5.16)"},{"comment":"The dimension-5 axial coupling ∂μσ Ψ̄γμγ5Ψ/Λσ is dismissed with a qualitative argument; a short quantitative estimate of the resulting non-Gaussianity would make it easier for the reader to verify that this channel indeed gives no substantial enhancement in the curvaton scenario.","section":"Sec. 4.3"},{"comment":"The manuscript contains several typos and formatting artifacts, such as 'non-renormalizabality' in Section 1 and an unclosed parenthesis in Eq. (5.8). A careful proofread would improve readability.","section":"General"}],"recommendation":"major_revision","confidential_remarks":"This is a well-written paper with explicit one-loop calculations and a clear physical message, and I do not see grounds for rejection. The central concern is whether the 4H benchmark is consistent with the paper's own control condition (5.3); if the authors can justify the mediator construction or move the benchmark to a compliant cutoff, the observability claim for loop-level fermion signals may need to be softened but the main enhancement mechanism would remain. The paper is within scope for JHEP and will be of interest to the cosmological-collider community."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Here's my take: this is a good paper with a genuinely useful idea. In the curvaton scenario, the primordial-fluctuation sector can have a much lower EFT cutoff than the inflaton sector, so heavy particles coupled to the curvaton can produce non-Gaussianities far above the standard-inflation estimates. The three explicit amplitudes (tree-level Higgs, scalar loop, fermion loop) are concrete, and the appendices give enough detail to reproduce them. I believe the central mechanism is sound.\n\nThe paper is honest about what it does and doesn't do. It acknowledges the independent nearby work [28], and the proof-of-principle UV completions (bi-axion alignment, brane sequestering) are clearly flagged as such. The comparison to the standard-inflation baseline in Sec. 4 is a useful corrective: without the curvaton trick, charged scalar and fermion loops are indeed unobservable.\n\nThe soft spot is the mediator benchmark that carries the fermion-loop claim. In Sec. 5, integrating out a mediator with MΣ=3H and μΣ=6H gives an effective cutoff Λσ,e≈4H, which is below the paper's own control bound Vσ^{1/4}≈7H. And the plotted heavy-particle masses extend up to 7H, above both MΣ and Λσ,e, so the effective operator is being used outside its domain of validity. If you rescale the fermion loop to a cutoff that actually respects the control condition, the amplitude drops to roughly the 10^-4 sensitivity edge, so 'observable loop-level fermion signal' is the least secure quantitative statement. The scalar loop survives the rescaling and remains well above threshold, so the main qualitative conclusion — that curvaton-mediated loop effects are observable — holds at least for scalars.\n\nMinor notes: the tree-level formula is quoted from [10] rather than re-derived, which is fine; and the curvaton's own local fNL=-5/4 provides a smooth background that the heavy-particle oscillations must be extracted from, as the paper notes.\n\nWho is this for? Cosmological collider phenomenologists and anyone working on curvaton models. It deserves a serious referee. I would send it out with a request to pay attention to the EFT-control question around the mediator benchmark and to ask the authors to state the observable claims for a range of Λσ respectful of Vσ^{1/4}.","headline":"Solid, useful paper: curvaton's separate low EFT cutoff genuinely boosts heavy-particle NG, but the fermion-loop observability claim rests on a benchmark outside the paper's own control bound.","tokens_in":22581,"tokens_out":6561,"would_cite":true,"duration_ms":58869,"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":"In the curvaton scenario, heavy scalars and fermions can leave observable non-Gaussianities, with benchmark loop amplitudes of $10^{-2}$–$10^{-1}$ for scalars and $10^{-3}$ for fermions.","keywords":["curvaton","primordial non-Gaussianity","cosmological collider physics","squeezed limit","effective field theory","heavy particles","inflation","non-Gaussianity"],"falsifier":"A future cosmic-variance-limited 21-cm or large-scale-structure measurement of the squeezed bispectrum reaching $\\sigma_{f_{\\rm NL}}\\sim 10^{-4}$ that finds neither the predicted non-analytic oscillations at $|f_{\\chi,\\mathrm{loop}}|\\sim 10^{-2}$–$10^{-1}$ (or $|f_{\\Psi,\\mathrm{loop}}|\\sim 10^{-3}$) nor the local $f_{\\rm NL}^{\\rm loc}=-5/4$ would falsify the benchmark curvaton-collider scenario.","tokens_in":21329,"feed_emoji":"🔭","tokens_out":11630,"duration_ms":104100,"temperature":0.7,"pith_summary":"The paper asks whether heavy particles with masses near the inflationary Hubble scale $H$ can leave detectable cosmological-collider signatures. In standard single-field inflation, shift symmetry and the high energy scale of the inflaton potential suppress the relevant couplings, so non-Gaussianities from such particles are often unobservable. The paper's claim is that in the curvaton scenario—where one field drives the expansion and a second light field seeds the fluctuations—the curvaton can carry its own much lower effective-field-theory cutoff, making couplings to heavy scalars and fermions far stronger while keeping the EFT controlled. For benchmark parameters the loop-level bispectrum amplitudes reach $\\sim 10^{-1}$ for charged scalars and $\\sim 10^{-3}$ for charged fermions, orders of magnitude above the standard-inflation predictions. The setup also predicts a robust local non-Gaussianity $f_{\\rm NL}^{\\rm loc}=-5/4$, testable by upcoming surveys.","feed_headline":"Curvaton lifts heavy-particle cosmic signals into view","feed_subtitle":"A low curvaton EFT cutoff makes loop-level non-Gaussianities of order 0.1 reachable by future surveys.","key_machinery":"The load-bearing object is a two-sector sequestered effective field theory in which the curvaton $\\sigma$ has its own cutoff $\\Lambda_\\sigma$ much smaller than the inflaton cutoff $\\Lambda_\\phi$. Shift-symmetric derivative operators, $\\Lambda_\\sigma^{-2}(\\partial\\sigma)^2\\chi^\\dagger\\chi$ for scalars and $\\Lambda_\\sigma^{-3}(\\partial\\sigma)^2\\bar\\Psi\\Psi$ for fermions, couple the curvaton to heavy states; a heavier mediator with coupling $\\mu\\Sigma\\chi^\\dagger\\chi$ is integrated out to give $\\Lambda_{\\sigma,\\mathrm{eff}}\\simeq M_\\Sigma^2\\Lambda_\\sigma/\\mu\\sim 4H$. The diagnostic signal is the non-analytic squeezed-limit three-point function, whose scaling $\\propto (k_3/k_1)^{3+2i\\mu}$ for scalars and $\\propto (k_3/k_1)^{4+2i\\tilde\\mu}$ for fermions is the on-shell fingerprint of a particle with mass $\\sim H$.","core_discovery":"The central claim is that primordial fluctuations sourced by a curvaton rather than by the inflaton open a much larger window for heavy-particle signatures. With the inflaton and curvaton in sequestered sectors, the hierarchy $\\Lambda_\\phi \\gtrsim V_{\\rm inf}^{1/4} > 250H \\gg \\Lambda_\\sigma \\gtrsim V_\\sigma^{1/4}\\sim 10H$ is consistent, and after integrating out a mediator the effective curvaton cutoff can fall to $\\Lambda_\\sigma\\sim 4H$. Couplings such as $\\Lambda_\\sigma^{-2}(\\partial\\sigma)^2\\chi^\\dagger\\chi$ then generate squeezed-limit bispectra whose non-analytic momentum dependence $(k_3/k_1)^\\Delta$ carries the heavy mass and spin. Explicit in-in calculations give $|f_{\\chi,\\mathrm{loop}}|\\sim 10^{-2}$–$10^{-1}$ and $|f_{\\Psi,\\mathrm{loop}}|\\sim 10^{-3}$ for $\\Lambda_\\sigma=4H$, $\\dot\\sigma_0=-H^2$, making loop-level cosmological collider signals observable in principle.","pith_inferences":["Extension: the same two-sector construction should boost four-point (trispectrum) signals even more aggressively, because a trispectrum can be generated without a $\\dot\\sigma_0$ insertion, allowing a lower $V_\\sigma$ and hence a lower $\\Lambda_\\sigma$.","Extension: if the sequestered sectors are realized in an extra dimension, Kaluza–Klein gravitons necessarily mediate between them; computing their squeezed bispectrum would test whether spin-2 heavy states also become observable.","Extension: the axial coupling $\\partial_\\mu\\sigma\\,\\bar\\Psi\\gamma^\\mu\\gamma^5\\Psi$ is set aside in this paper; a dedicated calculation could reveal whether it produces a chemical-potential enhancement analogous to the inflaton case."],"forward_implications":["For benchmark parameters, loop-level non-Gaussianity from charged scalars and fermions moves from $|f|\\sim 10^{-9}$ in the standard inflationary paradigm to $|f|\\sim 10^{-2}$–$10^{-1}$ and $\\sim 10^{-3}$, putting these targets within reach of future large-scale-structure and 21-cm experiments.","A detection in the squeezed limit would measure the exponent $\\mu=\\sqrt{m^2/H^2-9/4}$ or $\\tilde\\mu=m_\\Psi/H$ and the associated angular dependence, giving on-shell mass and spin information for particles far beyond terrestrial collider energies.","No classical fine-tuning is needed: the curvaton-induced mass shift $\\dot\\sigma_0^2/\\Lambda_\\sigma^2$ is small, unlike the inflaton-induced $\\dot\\varphi_0^2/\\Lambda_\\varphi^2$ contamination in standard inflation.","Independent of the heavy particles, the scenario predicts a local non-Gaussianity $f_{\\rm NL}^{\\rm loc}=-5/4$, a signature of the curvaton that upcoming surveys are expected to test.","The same coupling portal applies to Standard Model gauge-charged states such as $W$-boson loops, so the 'heavy-lifted' Standard Model signals become promising targets for future observations."],"supporting_citations":[{"why":"It supplies the in-in squeezed-limit machinery and the loop-to-tree decomposition used for all bispectrum computations.","marker":"[10]"},{"why":"It provides the standard-inflation tree-level heavy-Higgs amplitudes and parametric estimates that define the baseline in Sec. 4.","marker":"[18]"},{"why":"It gives the spinor two-point function that appendix B matches, and proposes a structurally similar Higgs-fluctuation cosmological collider.","marker":"[28]"},{"why":"It sets the Planck bounds on non-Gaussianity and on the Hubble-to-Planck ratio used to fix the benchmark scales.","marker":"[31]"},{"why":"It forecasts future large-scale-structure sensitivity $\\sigma_{f_{\\rm NL}}\\sim 1$ used to argue observability.","marker":"[32]"},{"why":"It forecasts cosmic-variance-limited 21-cm sensitivity $\\sigma_{f_{\\rm NL}}\\sim 10^{-4}$, the ultimate observational reach cited in the paper.","marker":"[35]"},{"why":"It introduces the curvaton as a source of the curvature perturbation.","marker":"[40]"},{"why":"It derives the conversion of curvaton fluctuations into the final curvature perturbation, $\\zeta=\\frac{2}{3}\\frac{\\delta\\sigma}{\\sigma_0}$.","marker":"[41]"},{"why":"It develops cosmological moduli/curvaton effects on the CMB that anchor the phenomenological history.","marker":"[42]"},{"why":"It derives the local non-Gaussianity $f_{\\rm NL}^{\\rm loc}=-5/4$ in the curvaton scenario.","marker":"[47]"}],"fun_headline_variants":["Curvaton amplifies heavy-particle non-Gaussianities to observability","Curvaton scenario makes heavy particles visible in cosmic signals","Heavy particle loops detectable via curvaton-driven non-Gaussianity","Curvaton EFT enables future detection of heavy particle loops","Curvaton boosts cosmological collider signals to detectable levels"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The enhancement depends on the assumption that the inflaton and curvaton can be sequestered into two sectors with independent EFT cutoffs, so that the curvaton cutoff can sit at a few times $H$ even though the curvaton's field value is far above it; if this separation has no ultraviolet completion, the predicted boost in non-Gaussianity disappears.","fun_headline_variants_meta":{"raw":{"variants":["Curvaton amplifies heavy-particle non-Gaussianities to observability","Curvaton scenario makes heavy particles visible in cosmic signals","Heavy particle loops detectable via curvaton-driven non-Gaussianity","Curvaton EFT enables future detection of heavy particle loops","Curvaton boosts cosmological collider signals to detectable levels"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000302,"raw_usage":{"total_tokens":1720,"prompt_tokens":904,"completion_tokens":816,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":520,"completion_tokens_details":{"reasoning_tokens":729}},"tokens_in":520,"tokens_out":816,"duration_ms":7309,"temperature":1.0,"reasoning_tokens":729,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T10:17:08.780103+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A future cosmic-variance-limited 21-cm or large-scale-structure measurement of the squeezed bispectrum reaching $\\sigma_{f_{\\rm NL}}\\sim 10^{-4}$ that finds neither the predicted non-analytic oscillations at $|f_{\\chi,\\mathrm{loop}}|\\sim 10^{-2}$–$10^{-1}$ (or $|f_{\\Psi,\\mathrm{loop}}|\\sim 10^{-3}$) nor the local $f_{\\rm NL}^{\\rm loc}=-5/4$ would falsify the benchmark curvaton-collider scenario.","supporting_citations":[],"review_version":1}