{"id":"4f254173-f901-4dd5-9e2c-4ed89946e9de","arxiv_id":"2411.11716","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":4.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":5,"one_line_summary":"A pseudoscalar glueball from a dark Yang-Mills sector, called a GALP, is shown to behave like an axion-like particle and to be a viable dark matter candidate with a predicted mass-coupling relation.","lead":"This paper argues that the pseudoscalar glueball of a dark Yang-Mills sector behaves like an axion-like particle, with tiny couplings to photons and nucleons set by a heavy mediator. If correct, these 'glueball ALPs' (GALPs) are viable dark matter candidates across a huge mass range, and the paper maps where existing and future searches could find them.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Eq. (22)'s 5/3 mass slope is set by the un-checked assumption that the portal fermion Psi was in thermal equilibrium and annihilated with the branching assumed in Sec. V; no rate calculation is provided for this load-bearing step.","rationale":"Good-faith reading: the paper is an EFT/phenomenology study that introduces GALPs and maps their parameter space, with the mass-coupling relation Eq. (22) as the main predictive result. For Eq. (22) to hold, the portal fermion's thermal history must connect the dark and visible temperatures through Eq. (21). That history is asserted in Sec. V with no rate calculation; this is the least secure link in the argument. The concern does not imply the paper is wrong: if the rate check passes, Eq. (22) is a valid toy-model prediction. The reader's conditional verdict already captures this, so I do not change it. I also note the paper's own caveats ('toy model', 'we leave a precision analysis for a later study') and the m_GALP = 6 Lambda lattice-identification issue, but the thermal-equilibrium premise is the load-bearing one. A concrete Boltzmann calculation or a direct branching-ratio check would settle it.","tokens_in":21456,"tokens_out":12419,"duration_ms":118279,"concrete_test":"Compute the thermalization and decoupling of Psi in the model defined by Eq. (11) with dark SU(3): solve the coupled Boltzmann equations for rho_gamma and rho_g~ from T >> M_Psi to T << M_Psi, including Psi Psi-bar -> gamma gamma, Psi Psi-bar -> gamma g~, and Psi Psi-bar -> g~ g~, for representative parameters (M_Psi = 10 TeV, epsilon = 10^{-4}, N = 3). If the resulting zeta_T^{-1} deviates from Eq. (21) by more than about 30%, or if Gamma_{Psi Psi-bar -> gamma gamma}/H < 1 at T ~ M_Psi for any point in Fig. 3, then the m^{5/3} slope and normalization of Eq. (22) are not robust. A simpler analytic check: compute the ratio Gamma(Psi Psi-bar -> gamma g~) / Gamma(Psi Psi-bar -> gamma gamma) including color and phase space factors; if it is not approximately epsilon^2 / 2, the temperature ratio in Eq. (20) is incorrect.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Sec. V derives the central mass-coupling relation Eq. (22) from three premises: (i) at T >> M_Psi the heavy portal fermion Psi is in thermal equilibrium with photons and has energy density rho_Psi ~ rho_gamma; (ii) at T ~ M_Psi annihilation into gamma gamma dominates, and the dark-gluon density is produced mainly by Psi Psi-bar -> gamma g~ with efficiency rho_g~ ~ (1/2) epsilon^2 rho_Psi; (iii) subsequent evolution is entropy-dominated by the visible sector, giving Eq. (21), zeta_T^{-1} proportional to sqrt(epsilon). None of these premises is checked against the actual rates. In particular, the equilibrium condition requires n_Psi <sigma v>_{Psi Psi-bar -> gamma gamma} > H at T ~ M_Psi; for large M_Psi this can fail, and if Psi was never in equilibrium the whole zeta_T^{-1} proportional to sqrt(epsilon) scaling, and therefore the m^{5/3} slope in Eq. (22), is not justified. Even if equilibrium holds, the branching ratio into dark gluons and the factor 1/2 are order-of-magnitude estimates; a full Boltzmann computation could change the normalization by O(1) and, if other channels (e.g. Psi Psi-bar -> g~ g~) are not negligible, even the exponent. The paper explicitly calls this a toy model ('we briefly discussed'), but Eq. (22) is presented as the model's prediction. The self-interaction-mass inconsistency (120 vs 180 MeV) and the m_GALP = 6 Lambda identification are secondary; the thermal-history step is the least supported input to the headline relation.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper considers a dark SU(N) Yang-Mills sector that confines into a tower of glueballs and studies the lightest scalar (0++) and pseudoscalar (0-+) glueballs, calling the pseudoscalar a Glueball ALP (GALP). It constructs an effective potential for H and A from the trace anomaly (Eq. (7)), reviews the glueball relic-density estimate Eq. (9), introduces a dimension-8 portal to photons and QCD gluons through a heavy fermion Psi (Eqs. (11) and (16)), and derives effective photon and nucleon couplings (Eqs. (13) and (19)). Using published ALP constraints, it plots allowed regions in the mass-coupling plane (Figs. 1-3). The central new prediction is Eq. (22), a mass-coupling relation g_GALPgamma ~ 10^{-21.1} GeV^{-1} kappa (m_GALP/keV)^{5/3} (M_Psi/GeV)^{-4} obtained by imposing that GALPs account for all dark matter, with Lambda_0 = 133 eV and m_GALP = 6 Lambda.","tokens_in":21681,"tokens_out":11175,"duration_ms":101402,"significance":"If the construction holds, GALPs are a concrete composite realization of axion-like dark matter with a testable mass-coupling line, and the paper's careful mapping of existing photon, supernova, and beam-dump constraints to this line is useful. The paper gives credit to earlier scalar-glueball studies and clearly states that GALPs do not solve the strong-CP problem (Sec. VI). Its strengths are the trace-anomaly-constrained EFT setup, the explicit operator basis for the dimension-8 portal, and the falsifiable prediction Eq. (22). The main caveat is that the normalization and slope of Eq. (22) rely on an uncomputed thermal history, as detailed below.","major_comments":[{"comment":"The headline relation Eq. (22) is derived from the assumptions in Eqs. (20)-(21): Psi is in thermal equilibrium with the photon bath at T >> M_Psi, Psi Psi-bar -> gamma gamma dominates its annihilation, and the dark-gluon density is generated by Psi Psi-bar -> gamma g-tilde with an efficiency factor 1/2. No rate or Boltzmann calculation is provided for any of these premises. If Psi was never in equilibrium, or if Psi Psi-bar -> g-tilde g-tilde (or other channels) contributes significantly, the zeta_T^{-1} proportional to sqrt(epsilon) scaling in Eq. (21) and therefore the m^{5/3} slope and normalization of Eq. (22) change. Since Eq. (22) is presented in Sec. V and used in Fig. 3 as the model's mass-coupling relation, this step needs either a concrete rate computation or an explicit statement that Eq. (22) is an illustrative toy-model estimate; the Conclusions' description of the portal as a toy model is in tension with the central role Eq. (22) plays in the paper.","section":"This concerns Section V, Eqs. (20)-(22)."},{"comment":"The identification m_GALP = 6 Lambda for the pseudoscalar glueball is used to convert the relic-density condition into Eq. (22), but the lattice and model references cited (Refs. [2,92]) primarily determine the lightest scalar 0++ glueball mass; no separate determination of the 0-+ mass ratio is presented. Because Eq. (9) describes the relic density of the lightest glueball and the pseudoscalar can decay through a -> eta eta (Eq. (A6)), the assumption that the pseudoscalar provides all of dark matter with m_GALP = 6 Lambda needs explicit justification.","section":"This concerns Section V, Eq. (9) and Eq. (22), and Figs. 1-3."},{"comment":"The paper assumes that the dimension-8 operator in Eq. (11) is the dominant dark-visible interaction without deriving the operator basis or showing that lower-dimensional operators (for example dimension-6 pure-gauge operators generated at one loop by a fermion charged under both dark SU(N) and electromagnetism) are absent or suppressed. If such operators exist, the (Lambda/M_Psi)^4 scaling of Eq. (13) and hence the whole Eq. (22) relation would not survive. At minimum this should be stated as a model assumption rather than presented as the portal.","section":"This concerns Section IV.B, Eq. (11)."}],"minor_comments":[{"comment":"The self-interaction lower bound on the GALP mass is inconsistent: the caption of Fig. 1 and the text of Sec. V use m_GALP >= 120 MeV, while Sec. IV.C quotes m_GALP >= 180 MeV and the stable window 180 MeV <= m_a <= 270 MeV; the figures and text should use one value.","section":"This concerns the caption of Fig. 1 and Secs. IV.C and V."},{"comment":"The prefactor is written as 10^{-(21.1 +/- 0.4)} GeV, but the coupling g_GALPgamma has dimension GeV^{-1}; the prefactor should presumably read 10^{-(21.1 +/- 0.4)} GeV^{-1}.","section":"This concerns Eq. (22)."},{"comment":"Equation (7) contains an arbitrary function f(A/H) and five uncomputed coefficients c0...c4; the appendix provides consistency conditions but no solution, so the EFT is not predictive beyond the O(1) parameter kappa unless kappa is explicitly treated as a free input.","section":"This concerns Eq. (7) and Appendix A."},{"comment":"The phrase derive from the first principles overstates the status of Eq. (7), which is constrained by the trace anomaly but still contains arbitrary functions and uncomputed coefficients.","section":"This concerns the Introduction."},{"comment":"The abstract's claim of viable GALP dark matter from sub-eV to the Planck scale is broader than what Fig. 3 supports for 100% dark matter, which requires m_GALP above roughly 120-180 MeV; sub-eV GALPs appear only in constraints that do not assume they are all of the dark matter.","section":"This concerns the Abstract and Sec. V."}],"recommendation":"major_revision","confidential_remarks":"The paper overlaps substantially with the same authors' companion paper Ref. [26] (GALPs! Composite heavy axion-like Dark Matter, 2408.14245), and Fig. 3 is explicitly taken from it; the editor should verify that the incremental contribution of this manuscript is clearly delineated and that there is no dual-submission issue. The central new formula Eq. (22) is promising but needs the thermal-history calculation or an explicit caveat before publication."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Worth a look if you care about heavy ALP dark matter. The genuinely new content here is the nucleon-coupling analysis (Fig. 2, Eq. 19), the compiled photon-constraint map (Fig. 1), and the appendix showing how the free coefficients of Eq. (7) could in principle be fixed by observables like the a→ηη decay rate. The core GALP concept, the dimension-8 photophilic operator, and Eq. (22) itself were introduced in the authors' own Ref. [26], and Fig. 3 is taken from there. That is not a fatal problem—the paper is transparent about it—but it means the headline claim is an extension, not a first derivation.\n\nThe EFT chain is internally consistent: trace-anomaly-constrained potential, dimension-8 portal, and the use of published ALP limits are careful. The authors also deserve credit for flagging the toy-model nature of the early-Universe production mechanism in the conclusions.\n\nThe soft spots are real but not disqualifying. The main one is the thermal-history step behind Eq. (22): the 5/3 mass slope comes from assuming the portal fermion Ψ was in equilibrium with photons at T ≫ M_Ψ and that Ψ̄Ψ→γ g̃ dominates the dark-gluon production, giving ζ_T^{-1} ∝ √ε. I checked the text: there is no rate calculation for equilibrium, no Boltzmann estimate, and no justification for the 1/2 efficiency factor. The paper openly calls it a toy model, but Eq. (22) is then presented as the model's prediction. A full Boltzmann computation could change the normalization by O(1) and, if Ψ̄Ψ→g̃g̃ is not negligible, could change the exponent itself. The second concern is m_GALP = 6Λ for the pseudoscalar: the lattice results cited mainly anchor the scalar 0++ mass, while the 0−+ is typically 1.3–1.5 times heavier; using 6Λ for the pseudoscalar neglects that. There are also two minor internal inconsistencies: the self-interaction bound is quoted as both 120 MeV and 180 MeV in different sections, and the gray exclusion regions in Fig. 1 are admittedly rough extrapolations.\n\nWho is this for? Phenomenologists working on composite dark matter or ALP searches; it gives a concrete, testable target for heavy ALPs beyond the QCD axion band. The central construction is not circular and the paper does not overclaim—it just inherits some fragility from the earlier paper. I would send it to a serious referee, mainly with the request that the thermal-history section be upgraded from a toy model to at least a rough rate check. That is the difference between Eq. (22) being a conjecture and being a prediction.","headline":"A readable, honest EFT phenomenology paper that extends the same authors' earlier GALP proposal with new constraints and an appendix, but the headline mass-coupling relation rests on an unchecked thermal-history assumption.","tokens_in":22481,"tokens_out":904,"would_cite":true,"duration_ms":11200,"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 paper claims that a pseudoscalar glueball from a dark Yang-Mills sector, called a GALP, is a viable axion-like dark matter candidate whose photon coupling follows a definite mass-coupling line when it makes up all of the dark matter.","keywords":["glueball dark matter","pseudoscalar glueball","axion-like particles","dark Yang-Mills sector","dimension-8 operator","mass-coupling relation","supernova bounds"],"falsifier":"A lattice computation of the SU(3) $0^{-+}$ glueball mass that disagrees with $m_{\\rm GALP}=6\\Lambda$ by more than the stated uncertainty would rescale the mass axis of the predicted line; a measurement of a stable dark-matter pseudoscalar whose photon coupling lies far off Eq. (22) for every $M_\\Psi$ would rule out the scenario.","tokens_in":21009,"feed_emoji":"🌌","tokens_out":12963,"duration_ms":82753,"temperature":0.7,"pith_summary":"This paper proposes that the lightest pseudoscalar glueball of a confining dark Yang-Mills sector behaves like an axion-like particle, called a GALP, and can account for all of the dark matter. The authors derive the effective low-energy theory of scalar and pseudoscalar glueballs, and show that a dimension-8 gauge portal coupling dark gluons to photons and QCD gluons becomes, after confinement, a set of very weak axion-like couplings. Because the dark sector passes through a first-order confinement-deconfinement transition, the relic abundance depends mainly on the dark confinement scale; combining that with the temperature ratio set by a heavy mediator produces a definite mass-coupling relation for GALP dark matter. If the relation holds, GALPs occupy a largely unexplored corner of the axion-photon parameter space: masses above roughly 120 MeV, couplings far weaker than canonical QCD axions, and effective Peccei-Quinn scales that can be super-Planckian. The paper therefore gives concrete search targets for photon-line, supernova, and beam-dump experiments.","feed_headline":"Dark glueballs can be axion-like dark matter","feed_subtitle":"A predicted mass-coupling line puts heavy, weakly coupled GALPs in reach of supernova, beam-dump, and photon searches.","key_machinery":"The central object is the GALP, a pseudoscalar glueball $0^{-+}$ of a confining dark SU(N) sector, whose mass is tied to the dark confinement scale by $m_{\\rm GALP}\\simeq 6\\Lambda$. Three ingredients carry the argument: the dimension-8 gauge portal of Eq. (11), which couples dark gluons to photons with strength $\\epsilon^2\\kappa(\\Lambda/M_\\Psi)^4$ after confinement and similarly couples dark gluons to QCD gluons; the glueball effective potential of Eq. (7), constrained by the dilatation anomaly, which fixes the spectrum and the order-one normalization $\\kappa$; and the relic-abundance formula of Eq. (9) together with the temperature ratio of Eq. (21), which converts the confinement scale into the observed dark-matter density and, with Eq. (13), into the mass-coupling relation of Eq. (22). The role of $\\kappa$ is to absorb the matrix-element normalizations from the glueball effective theory, so the prediction is parametric in the mediator mass $M_\\Psi$ and the small coupling $\\epsilon$.","core_discovery":"The central claim is that a dark SU(3) Yang-Mills sector whose only visible interaction is a dimension-8 operator generated by a heavy fermion $\\Psi$ produces a pseudoscalar glueball that is simultaneously a dark-matter candidate and an axion-like particle. The paper derives the glueball effective potential and identifies the pseudoscalar mass with the confinement scale, $m_{\\rm GALP}\\simeq 6\\Lambda$ for $N=3$. Assuming GALPs constitute all of the dark matter, the relic-density estimate combined with the dark-to-visible temperature ratio $\\zeta_T^{-1}\\propto\\sqrt{\\epsilon}$ gives $$g_{\\rm GALP\\gamma}\\simeq $10^{{-(21.1\\pm 0.4)}}$\\,\\mathrm{GeV}^{-1}\\,\\kappa\\,(m_{\\rm GALP}/\\mathrm{keV})^{5/3}(M_\\Psi/\\mathrm{GeV})^{-4},$$ with $\\kappa\\sim 1$ expected. The same effective theory generates GALP-nucleon couplings of order $g_{\\rm GALP p}\\sim -4.4\\times10^{-22}\\,\\epsilon^2\\kappa\\,(\\Lambda/100\\,\\mathrm{eV})^3(M_\\Psi/\\mathrm{GeV})^{-4}$. The paper maps the resulting parameter space against laboratory, stellar, supernova, and cosmological bounds and identifies a window around $180$ to $270$ MeV where a GALP is stable against decays and can be all of the dark matter.","pith_inferences":["The $5/3$ slope in Eq. (22) is a fingerprint of the dimension-8 portal together with the temperature-ratio scaling; a dimension-6 portal would give different powers of $M_\\Psi$ and a different line, so measuring the slope would identify the portal's operator dimension.","A dedicated calculation of $\\Psi\\bar{\\Psi}$ annihilation into a dark-gluon pair could sharpen the normalization of $\\zeta_T^{-1}$; the current relation assumes that channel is subleading by an $\\epsilon^2$ factor.","The model does not address the strong-CP problem, so discovering GALPs would not validate composite-axion solutions; conversely, GALP dark matter and QCD-axion dark matter could coexist and be distinguished by their very different two-photon line energies.","The same glueball effective theory predicts a scalar partner with scalar and pseudoscalar photon couplings, so measuring the angular and polarization structure of a $\\gamma\\gamma$ line could tell GALP dark matter apart from a single axion-like particle."],"forward_implications":["If GALPs are all of the dark matter, their photon coupling is fixed by Eq. (22) once the GALP mass and mediator mass $M_\\Psi$ are chosen, so photon-line, supernova, and beam-dump searches have a definite target curve.","Because the dark matter is produced by confinement rather than freeze-out, GALP masses can sit far above the usual unitarity bound, up to the mediator scale, opening a heavy-axion window that has been largely unexplored.","The same effective potential gives a scalar glueball partner with comparable couplings, so a signal with two glueball-like states would point to a two-component dark-matter sector rather than a single ALP.","The effective Peccei-Quinn scale $f_a$ can be super-Planckian while the underlying new physics sits at TeV-PeV scales, populating a parameter region ordinarily considered inaccessible and motivating searches there.","Nucleophilic GALPs have a narrow stability window near 180 to 270 MeV, part of which is already constrained by SN 1987A cooling and Kamiokande-II; sharper supernova observations would test the rest."],"supporting_citations":[{"why":"Supplies the scalar glueball relic-density calculation summarized in Eq. (9) and the bounds on the confinement scale quoted in Eq. (10).","marker":"[1, 2]"},{"why":"Source of the dimension-8 dark-gluon-photon operator of Eq. (11) that becomes the GALP-photon coupling after confinement.","marker":"[4]"},{"why":"Earlier paper by the same authors that introduced the pseudoscalar glueball GALP and the mass-coupling relation that this work extends.","marker":"[26]"},{"why":"Lattice determination that $g/\\beta(g)\\sim 1$ at confinement, fixing the normalization of the glueball-photon coupling in Eq. (13).","marker":"[62]"},{"why":"Provides the BBN decay bound that sets the upper edge of the allowed decaying-GALP dark-matter region.","marker":"[84]"},{"why":"Supplies the chiral perturbation theory result converting the GALP-gluon coupling into the GALP-nucleon coupling of Eq. (19) and the QCD-axion comparison band.","marker":"[88]"},{"why":"Lattice simulations relating glueball masses to the confinement scale, supporting the $m_{\\rm GALP}=6\\Lambda$ identification used in Eq. (22).","marker":"[92]"},{"why":"Self-interaction bound $\\sigma<1/m^2$ used to set the lower limit on the GALP dark-matter mass.","marker":"[93]"}],"fun_headline_variants":["Dark glueballs as axion-like dark matter","Glueball axion-like particles could be dark matter","Dark glueballs mimic axion-like signals","GALPs: composite axion-like dark matter","Pseudoscalar glueballs: new dark matter axions"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The prediction rests on the dark and visible sectors communicating only through the dimension-8 operator generated by one heavy fermion, and on that fermion having been in thermal equilibrium with photons; if any other portal exists or the annihilation history differs, Eq. (22) changes.","fun_headline_variants_meta":{"raw":{"variants":["Dark glueballs as axion-like dark matter","Glueball axion-like particles could be dark matter","Dark glueballs mimic axion-like signals","GALPs: composite axion-like dark matter","Pseudoscalar glueballs: new dark matter axions"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000429,"raw_usage":{"total_tokens":2235,"prompt_tokens":1027,"completion_tokens":1208,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":643,"completion_tokens_details":{"reasoning_tokens":1132}},"tokens_in":643,"tokens_out":1208,"duration_ms":10212,"temperature":1.0,"reasoning_tokens":1132,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T18:16:00.343095+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A lattice computation of the SU(3) $0^{-+}$ glueball mass that disagrees with $m_{\\rm GALP}=6\\Lambda$ by more than the stated uncertainty would rescale the mass axis of the predicted line; a measurement of a stable dark-matter pseudoscalar whose photon coupling lies far off Eq. (22) for every $M_\\Psi$ would rule out the scenario.","supporting_citations":[],"review_version":1}