{"id":"d13219b6-3614-4651-a858-c10ed688823a","arxiv_id":"2501.09939","paper_version":1,"verdict":"REJECT","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"high","formal_verification":"none","parameter_count":0,"one_line_summary":"The claimed infrared slope for isocurvature-induced gravitational waves is 3 minus 4 divided by the logarithm of the effective peak scale squared over six k squared, approaching 3 in the deep infrared.","lead":"This paper derives an infrared slope formula for gravitational waves induced by isocurvature density perturbations, comparing it with the adiabatic case. The result is meant as a way to tell the two perturbation modes apart using future gravitational wave detectors.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Equation (16) omits the u^2v^2 factor from the substitution of P(k) = (k/k_eq)^2 P~(k), changing the infrared kernel from O(1) to O(v^{-4}); the slope formula (23) therefore does not follow, and Eq. (22) has a sign/reciprocal typo.","rationale":"The reader's weakest_assumption identifies the same algebraic error: the transformation from Eq. (10) to Eq. (15) is invalid because Eq. (16) omits the u^2v^2 factor that arises from P(k) = (k/k_eq)^2 P~(k). My independent check confirms the magnitude of the error: with the missing factor, the diagonal kernel at large v scales as v^{-4}; with the correct factor it is O(1) (times logarithms). This changes the power of k in the infrared integral, so Eq. (23) cannot be derived from the stated equations. I also verified the separate inconsistency between Eq. (22) and Eq. (23): taking the logarithmic derivative of Ω_GW ∝ (\\tilde k_*/k)^3 ln^2(...) gives -3 - 4/L, not 3 - 4/L. Since the numerical checks in Figs. 1-3 use the same approximated kernel, they validate the mistake rather than the physics. No other load-bearing concern is needed; this one alone invalidates the central claim. The appropriate verdict remains REJECT, agreeing with the reader's assessment.","tokens_in":8546,"tokens_out":15898,"duration_ms":135231,"concrete_test":"Recompute the infrared slope numerically from the unapproximated original expression: use Eq. (10) with Eq. (11) and a lognormal spectrum P(k) = (k/k_eq)^2 P~(k), with P~(k) as in Eq. (29), σ = 0.2, and k_eq/k_* = 10^3. Evaluate n_GW = d lnΩ_GW/d ln k at k = 10^{-4} \\tilde k_*. Compare this direct result against Eq. (23). If the direct numerical slope differs from the predicted value by more than 0.1, the paper's central claim is refuted.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central result, Eq. (23), depends on the compact form Eq. (15) with kernel Eq. (16). Substituting P(k) = (k/k_eq)^2 P~(k) into Eq. (10) gives P(vk)P(uk) = u^2v^2(k/k_eq)^4 P~(vk)P~(uk). Multiplying the explicit (k_eq/k)^4 in Eq. (11) yields a new kernel \\tilde I(u,v) = (u^2v^2/24) [4v^2-(1+v^2-u^2)^2]^2/(16u^2v^2) (I_c^2+I_s^2). Equation (16) drops the u^2v^2 factor. This is not a harmless rescaling. At u≈v>>1, the retained kernel behaves as \\tilde I(v,v) ≈ (27/64) ln^2(v^2/6)/v^4, whereas the correct kernel is asymptotically constant, \\tilde I(v,v) ≈ (81/384)[π^2+ln^2(v^2/3)]. The spurious v^{-4} is exactly what converts the narrow-peak v-integral into the (\\tilde k_*/k)^3 factor in Eq. (22); with the correct kernel the infrared scaling changes qualitatively. Moreover, Eq. (22) states Ω_GW ∝ (\\tilde k_*/k)^3 ln^2(\\tilde k_*^2/(6k^2)), whose logarithmic derivative is -3 - 4/ln(...), not the 3 - 4/ln(...) claimed in Eq. (23). Thus the isocurvature slope and the adiabatic/isocurvature discriminant are not consequences of the paper's own equations.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This manuscript studies the infrared behavior of gravitational waves induced by isocurvature scalar perturbations. The authors define an effective scalar power spectrum by absorbing the factor (k_eq/k)^2 into P(k), rewrite the GW energy density in a compact kernel form, and derive a log-dependent infrared spectral slope n_GW = 3 - 4/ln(k~*^2/(6k^2)). They compare this slope with the adiabatic case and conclude that the two cases are observationally distinguishable. The central claim is the slope formula in Eq. (23), together with the asserted robustness of the isocurvature/adiabatic discriminant.","tokens_in":8899,"tokens_out":9157,"duration_ms":84527,"significance":"If established, this result would extend the known infrared-slope analysis of induced GWs to isocurvature perturbations and could provide a useful observational discriminant. The paper builds on an existing kernel from the literature and gives explicit analytic expressions, which is helpful for checking the derivation. However, the central derivation contains a load-bearing algebraic error in the kernel reduction, and an additional inconsistency between Eqs. (22) and (23) means that the claimed slope formula does not follow from the paper's own equations. The topic is timely, but the present version does not support its main conclusion.","major_comments":[{"comment":"Substituting the effective spectrum eP(k) = (k_eq/k)^2 P(k) into Eq. (10) and using Eq. (11) yields an integrand proportional to u^2 v^2 (k_eq/k)^4 eP(uk)eP(vk). The k-independent kernel in Eq. (15) must therefore contain a factor u^2 v^2; explicitly, it should be (u^2 v^2/24)[4v^2-(1+v^2-u^2)^2]^2/(16u^2v^2)(I_c^2+I_s^2). Equation (16) omits this u^2 v^2 factor. This is not a harmless rescaling: the omitted factor changes the asymptotic behavior of the kernel at large u and v, and Eqs. (19)-(21), which feed directly into the scaling result Eq. (22), are derived from the incorrect kernel. Consequently the central slope formula Eq. (23) does not follow from the stated equations.","section":"Eqs. (14)-(16)"},{"comment":"Equation (22) states Omega_GW(k) is proportional to (k~_*/k)^3 ln^2(k~_*^2/(6k^2)). Taking the logarithmic derivative gives d ln Omega_GW/d ln k = -3 - 4/ln(k~_*^2/(6k^2)), not the value 3 - 4/ln(...) claimed in Eq. (23). If the intended scaling is (k/k~_*)^3 rather than (k~_*/k)^3, then Eq. (23) could be consistent, but Eq. (22) would have to be corrected. As written, Eq. (23) is not the derivative of Eq. (22).","section":"Eq. (22) versus Eq. (23)"},{"comment":"The refined near-peak kernel in Eq. (25) and the resulting expressions for Omega_GW and n_GW in Eqs. (26)-(27) are all obtained from the same incorrect kernel \tilde I(u,v) of Eq. (16). The error therefore propagates into the near-peak analysis as well. Before these results can be used, the calculation must be redone with the correctly reduced kernel, and the numerical agreement claimed in Figs. 1-3 must be reassessed.","section":"Sec. III.B, Eqs. (25)-(27)"}],"minor_comments":[{"comment":"Equation (34) is stated for 0 < alpha <= 1, but Fig. 4 indicates the shaded range is for 0 <= alpha <= 1; at alpha = 0 the argument of the logarithm diverges, so the domain used in the figure should be specified consistently.","section":"Eq. (34) and Fig. 4"},{"comment":"The typeset form of Eq. (27) is garbled in places; the term involving '22/33k2' should be checked and rewritten so that the formula can be verified.","section":"Eq. (27)"},{"comment":"The factor preceding the bracket in Eq. (6) is typeset ambiguously; please clarify the intended expression, e.g. (3 S_k / (2 sqrt(2))) (k_eq / k) [ ... ].","section":"Eq. (6)"},{"comment":"The tilde over P in eP(k) is used from Eq. (14) onward but is easy to confuse with the original P(k); consider using a distinct symbol such as P_eff(k) for clarity.","section":"Notation"}],"recommendation":"reject","confidential_remarks":"The algebraic error in the reduction from Eq. (10) to Eq. (16) is substantive and invalidates the central claim; the IR analysis would need to be redone from scratch, and it is not clear that the claimed slope survives. The paper may be worth a new submission if the corrected kernel still yields a qualitatively similar result, but that is not demonstrated here. I also note that the numerical agreement claimed in Figs. 1-3 is surprising given the algebra error, and the authors should be asked to check whether those numerics inadvertently used the same incorrect kernel."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The topic is real and the question is worth asking: does the IR slope of the induced GW spectrum distinguish adiabatic from isocurvature scalar perturbations? The paper is clearly written and follows the standard Yuan et al. framework, applying it to Domenech's isocurvature kernel. That's a legitimate extension, not a crank submission.\n\nBut the central derivation doesn't hold. Absorbing the k-dependence into an effective power spectrum in Eq. (14) is fine, but when they substitute P(vk)P(uk) into Eq. (10), they forget the u^2 v^2 factor that comes with the substitution. The correct effective kernel is Eq. (16) multiplied by u^2 v^2. Eq. (16) as written is just the original kernel with the (k_eq/k)^4 stripped off, which is not what you get after integrating out the k-dependence. At large u≈v this makes the kernel decay as v^{-4} instead of going to a constant. That spurious v^{-4} is exactly what produces the (k~*/k)^3 scaling in Eq. (22) and the claimed slope. With the correct kernel the IR integral scales differently, and the paper's central discriminant is not established.\n\nThere's also a separate sign/reciprocal problem: Eq. (22) writes Ω ∝ (k~*/k)^3, whose logarithmic derivative is -3 - 4/ln(...), not the 3 - 4/ln(...) in Eq. (23). So the paper's own equations don't even agree with each other. The numerical checks use the same approximate kernel, so they just reproduce the error.\n\nMy verdict is reject. The idea is fine and the paper is readable, but the load-bearing result is an artifact of an algebraic slip. I wouldn't send this to review as is; if the authors correct the substitution and recompute, they might find a different IR behavior worth reporting. But that's a new calculation, not a patch.","headline":"A clearly written attempt to extend the IR-slope framework to isocurvature IGWs, but the main result rests on a missed u^2v^2 factor; the claimed discriminant doesn't follow.","tokens_in":9471,"tokens_out":16100,"would_cite":false,"duration_ms":140347,"reading_group":"no","serious_thinker":"no","would_accept_peer_review":false},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["04.30.-w","98.80.-k"],"model":"deepseek-v4-flash","headline":"The paper derives a log-dependent infrared slope for gravitational-wave spectra from isocurvature perturbations and argues it cleanly distinguishes them from adiabatic modes.","keywords":["induced gravitational waves","isocurvature perturbations","infrared spectral slope","primordial power spectrum","stochastic gravitational wave background","lognormal spectrum","broken power spectrum","radiation domination"],"falsifier":"A direct numerical evaluation of Eq. (10) using the original unrescaled kernel of Eq. (11) with a lognormal spectrum, for $k\\ll k_*$, would settle the claim: if the numerically obtained slope departs from $3-4/\\ln(\\tilde{k}_*^2/6k^2)$, the rescaling assumption fails. An even quicker check is symbolic algebra: substitute $P(k)=(k/k_{\\rm eq})^2\\tilde{P}(k)$ into Eq. (11) and count powers of $u$ and $v$ in the resulting kernel to see whether a $1/(u^2v^2)$ factor survives.","tokens_in":8277,"feed_emoji":"🌊","tokens_out":7622,"duration_ms":74357,"temperature":0.7,"pith_summary":"The paper predicts the low-frequency (infrared) shape of the gravitational-wave background produced when isocurvature perturbations—relative fluctuations between matter and radiation with no overall density perturbation—source tensor perturbations at second order. Its central result is that the spectral slope obeys $n_{\\rm GW}=3-4/\\ln(\\tilde{k}_*^2/6k^2)$, where $\\tilde{k}_*$ is the peak scale of the rescaled scalar power spectrum. In the deep infrared the slope approaches 3 for any spectrum shape, but at observable intermediate scales it lies measurably below the adiabatic prediction $3-4/\\ln(4k_*^2/3k^2)$. A separate near-peak expression is derived for narrow spectra. This matters because small-scale isocurvature modes are almost unconstrained, and a slope measurement is a direct way to identify their presence.","feed_headline":"Gravitational-wave slope unmasks isocurvature perturbations","feed_subtitle":"The predicted infrared index drops by a scale-dependent logarithmic term, cleanly separating two primordial origins.","key_machinery":"The load-bearing device is the rescaling of the primordial spectrum, $\\tilde{P}(k)=(k_{\\rm eq}/k)^2P(k)$, which absorbs the explicit $k$ dependence of the isocurvature integration kernel and leaves a formally $k$-independent kernel $\\tilde{I}(u,v)$. In the infrared limit $u,v\\gg 1$ this kernel reduces to the closed form Eq. (19), whose diagonal value $\\tilde{I}(v,v)\\simeq 27\\ln^2(v^2/6)/(64v^4)$ supplies the logarithmic factor in the slope. The effective peak scale $\\tilde{k}_*$ of $\\tilde{P}$ then fixes the upper endpoint of the dominant $v$-integral, and a small-$y$ expansion around $y=vk/\\tilde{k}_*-1$ produces the scaling $\\Omega_{\\rm GW}\\propto(\\tilde{k}_*/k)^3\\ln^2(\\tilde{k}_*^2/6k^2)$, from which Eq. (23) follows.","core_discovery":"The paper claims that gravitational waves induced by isocurvature scalar perturbations during radiation domination have an infrared spectral slope that is logarithmic in the ratio of the effective peak scale to the observation scale. Concretely, $n_{\\rm GW} \\equiv d\\ln\\Omega_{\\rm GW}/d\\ln k = 3 - 4/\\ln(\\tilde{k}_*^2/6k^2)$, with $\\tilde{k}_*$ the peak of $\\tilde{P}(k)=(k_{\\rm eq}/k)^2P(k)$. In the limit $k/\\tilde{k}_*\\to 0$ the slope universally approaches 3, while the adiabatic analogue has the same universal limit but a different logarithmic correction, $3-4/\\ln(4k_*^2/3k^2)$. The paper further shows analytically and numerically that for narrow spectra the near-peak slope is independent of the functional form of the spectrum, and that the $\\Omega_{\\rm GW}$ peak is shifted toward smaller $k$ relative to the matter power spectrum peak.","pith_inferences":["I infer that the discriminating power is strongest in the intermediate infrared, roughly a few decades below the peak, because both formulas converge to 3 at arbitrarily small $k$; the paper's figures show the gap but do not state this emphasis.","The same effective-spectrum resummation could be tried for Poisson-type isocurvature sources such as primordial black hole clustering, whose matter spectrum would give a different $\\tilde{k}_*$; the paper does not treat that case.","A testable extension is to compare the predicted slope against the full one-loop tensor power spectrum computed without the narrow-width approximation on wide spectra; agreement there would strengthen confidence in Eq. (23)."],"forward_implications":["In the deep infrared, the isocurvature-induced gravitational-wave slope is universally 3, independent of the shape of the scalar power spectrum.","For lognormal spectra the effective peak scale is $\\tilde{k}_*=e^{-2\\sigma^2}k_*$; for broken-power spectra it is $\\tilde{k}_*=\\min\\{2k_-,k_*\\}$, so the gravitational-wave peak sits at lower frequencies than the matter peak.","The predicted separation between isocurvature and adiabatic slope curves gives pulsar timing arrays and space-based interferometers a concrete spectral index to test.","For narrow spectra the near-peak slope expression does not depend on the specific form of the spectrum and improves the match to numerical integration."],"supporting_citations":[{"why":"Supplies the log-dependent infrared slope framework for adiabatic-induced gravitational waves that this paper adapts and compares against.","marker":"[26]"},{"why":"Establishes the universal infrared slope $n_{\\rm GW}\\to 3$ that the isocurvature formula reproduces in the deep infrared.","marker":"[27]"},{"why":"Provides the isocurvature perturbation solution and the kernel $I(u,v,k)$ used as the starting point of the calculation.","marker":"[28]"},{"why":"Gives the second-order gravitational-wave formalism from primordial scalar perturbations.","marker":"[3]"},{"why":"Derives the induced gravitational-wave spectrum from scalar perturbations, the basis of Eq. (10).","marker":"[4]"},{"why":"Motivates the observational goal by asking whether pulsar timing arrays can distinguish adiabatic from isocurvature fluctuations.","marker":"[25]"},{"why":"Supplies the lognormal and broken-power-spectrum models used to test the analytical slopes numerically.","marker":"[29]"}],"fun_headline_variants":["Isocurvature GWs get a log-dependent infrared slope","Infrared GW slope separates isocurvature from adiabatic","Log term in GW slope reveals isocurvature perturbations","A logarithmic twist in GW slope pinpoints isocurvature","Infrared slope of induced GWs flags isocurvature origin"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The derivation assumes the step from Eq. (10) to Eq. (15) is exact: pulling $(k_{\\rm eq}/k)^2$ out of the kernel and into the power spectrum must leave a kernel whose large-$u,v$ behavior is exactly that of Eq. (19); if any residual inverse power of $u$ or $v$ remains, the logarithmic slope formula no longer follows from the stated equations.","fun_headline_variants_meta":{"raw":{"variants":["Isocurvature GWs get a log-dependent infrared slope","Infrared GW slope separates isocurvature from adiabatic","Log term in GW slope reveals isocurvature perturbations","A logarithmic twist in GW slope pinpoints isocurvature","Infrared slope of induced GWs flags isocurvature origin"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00026,"raw_usage":{"total_tokens":1561,"prompt_tokens":890,"completion_tokens":671,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":506,"completion_tokens_details":{"reasoning_tokens":585}},"tokens_in":506,"tokens_out":671,"duration_ms":6738,"temperature":1.0,"reasoning_tokens":585,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T19:33:49.312269+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A direct numerical evaluation of Eq. (10) using the original unrescaled kernel of Eq. (11) with a lognormal spectrum, for $k\\ll k_*$, would settle the claim: if the numerically obtained slope departs from $3-4/\\ln(\\tilde{k}_*^2/6k^2)$, the rescaling assumption fails. An even quicker check is symbolic algebra: substitute $P(k)=(k/k_{\\rm eq})^2\\tilde{P}(k)$ into Eq. (11) and count powers of $u$ and $v$ in the resulting kernel to see whether a $1/(u^2v^2)$ factor survives.","supporting_citations":[],"review_version":1}