{"id":"149bcd59-aa05-4cb1-98e4-bc049d3f2366","arxiv_id":"2507.05353","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":5,"one_line_summary":"A temporary axion mass spike during a first-order electroweak phase transition can make the QCD axion dark matter over a wide range of decay constants without a tuned initial angle.","lead":"This paper proposes a new axion dark matter production stage, called recurrent misalignment, that uses a first-order electroweak phase transition to temporarily give the QCD axion an extra mass. If correct, the QCD axion could match dark matter abundance for decay constants from 1e8 to 1e14 GeV without tuning the initial misalignment angle.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Eq. (11) uses a period-2π sine force, but the S^4Φ^2 operator in Eq. (1) gives a period-π axion potential (∝sin2θ); the reported relic curves are for a different model.","rationale":"The reader's weakest_assumption (spatial patchiness and the abrupt homogeneous v_s drop) is a legitimate idealization issue, but I think the single most load-bearing concern is more basic: the numerical solver uses the standard period-2π cosine equation of motion, whereas the explicit S^4Φ^2 operator in Eq. (1) generates a period-π axion potential. This is an internal consistency problem, not a disagreement with external consensus. It directly affects the evolution of θ during the crucial first oscillation phase, the phase drift below Tn, and therefore the final relic density for each (f_a,Λ). The concrete test is a one-line change to the force term, so the authors can likely fix it; the recurrent-misalignment idea is not ruled out by this concern alone. For that reason I keep the reader's CONDITIONAL verdict rather than moving to REJECT. I disagree with the reader that the homogeneous-vs-bubble idealization is the weakest point: even in the idealized homogeneous limit, Eq. (11) is not the equation of motion of the operator in Eq. (1), and the reported Λ values in Fig. 3 are not yet evidence for the stated model. If the corrected equation reproduces a broad f_a range, the spatial-patchiness concern would then become the next important check, but it is secondary to the EOM mismatch.","tokens_in":15493,"tokens_out":20267,"duration_ms":262605,"concrete_test":"Replace the S-induced term in Eq. (11) by the exact force: θ¨+3Hθ˙+m_a0^2(T)sinθ + [2v_s^4(T)/Λ^2] sin(2θ)=0 (equivalently (m_as^2/2)sin2θ), with the same θ_i, Ts, Tn, BP1 and the same switch-off at Tn. Recompute Ω_a h^2 for a grid of (f_a,Λ) spanning f_a=10^8–10^14 GeV and compare with the published curves. If the corrected equation still yields a Λ∈(f_a,M_Pl) with Ωh^2=0.12 for every f_a in the range, the concern is resolved; if the required Λ shifts significantly or such Λ ceases to exist for part of the range, the central claim is not supported by the current numerics.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The most load-bearing flaw is an internal inconsistency between Eq. (1) and Eq. (11). Setting S→v_s and Φ≃(f_a/√2)e^{iθ}, the operator S^4 Φ^2 e^{iα}/Λ^2 + h.c. in Eq. (1) produces an axion potential proportional to cos(2θ+α), not to cosθ. With α=π this is −cos(2θ), with minima at θ=0 and θ=π and an equation-of-motion force ∝ sin(2θ); the curvature at θ=0 is m_as^2=4v_s^4/Λ^2, but the finite-amplitude force is (m_as^2/2)sin(2θ), not m_as^2 sinθ. The paper instead inserts m_as^2 into a single mass-squared in Eq. (2) and solves Eq. (11), θ¨+3Hθ˙+m_a^2(T)sinθ=0, which is the equation for a single period-2π cosine potential. The two equations differ already at θ_i∼1 (e.g., at θ=1, sin2θ/2≈0.46 vs sin1≈0.84) and differ qualitatively for θ_i>π/2, where the true S-potential has a local minimum at π. Because every Ω_a h^2 point in Fig. 3 is obtained from Eq. (11), the numerical support for f_a∈[10^8,10^14] GeV does not yet correspond to the model defined by Eq. (1). This concern is prior to the bubble-patchiness issue: even in the idealized global-abrupt v_s limit, the dynamics solved are not the dynamics of the stated operator.","agreement_with_reader":"disagree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes a 'recurrent misalignment' mechanism for the QCD axion: a first-order electroweak phase transition is induced by a Z2-odd real scalar S, and an explicit PQ-breaking dimension-6 operator S^4 Phi^2 e^{i alpha}/Lambda^2 gives the axion a temporary additional mass while <S> is nonzero. The axion is claimed to undergo an early oscillation phase between T_s and T_n, then to drift in field space when <S> drops to zero, and finally to resume misalignment oscillations under the QCD potential; by choosing Lambda for each f_a, the authors claim Omega_a h^2 = 0.12 can be obtained for f_a in [10^8, 10^14] GeV with O(1) theta_i. The paper also computes the gravitational-wave signal of the FOEWPT for two benchmark points.","tokens_in":15955,"tokens_out":12121,"duration_ms":140981,"significance":"The intended result is significant: it would convert the usual single-scale prediction f_a ~ 10^12 GeV into a broad experimentally accessible window and tie axion DM to electroweak-scale gravitational-wave probes. The paper is clearly written, gives a UV-completion sketch for the dimension-6 operator, specifies two benchmark points for the phase transition, and reports numerical results obtained with standard tools (FindBounce, AxionLimits). The central quantitative claim is, however, compromised by an inconsistency between the Lagrangian operator in Eq. (1) and the equation of motion actually solved in Eq. (11), so the present numerical support does not yet demonstrate the stated model's relic window.","major_comments":[{"comment":"The equation solved in Eq. (11), theta_ddot + 3H theta_dot + m_a^2(T) sin theta = 0, is the equation of motion for a single period-2pi potential V proportional to (1 - cos theta). Substituting Phi = (f_a/sqrt2) e^{i theta} and alpha = pi into the operator S^4 Phi^2 e^{i alpha}/Lambda^2 + h.c. in Eq. (1) gives V_extra = - (v_s^4 f_a^2 / Lambda^2) cos(2 theta), whose force is (m_as^2/2) sin(2 theta) with m_as^2 = 4 v_s^4 / Lambda^2 as in Eq. (2). This force differs from m_as^2 sin theta by the factor cos theta: it is reduced by about 0.54 at theta = 1 and changes sign for theta > pi/2, where the extra potential has a minimum at theta = pi and repels the field from theta = 0. Since every relic point in Fig. 3 is obtained by integrating Eq. (11), the numerical support for f_a in [10^8, 10^14] GeV does not yet describe the model defined by Eq. (1). The authors should solve the actual potential (QCD potential plus the cos(2 theta + alpha) term) or, alternatively, replace the operator so that the intended period-2pi potential is generated.","section":"Eq. (1) versus Eq. (11)"},{"comment":"The three-stage mass prescription in Eq. (2) treats v_s(T) as a globally homogeneous background that switches off discontinuously at T_n. A first-order electroweak phase transition proceeds by bubble nucleation, so during the transition the axion field in false-vacuum and true-vacuum regions sees different masses; the duration of the transition and the percolation details determine how much the field drifts before the second oscillation starts at T_osc^II. The claimed relic range is obtained from a single homogeneous theta(t) evolved through this abrupt drop, so the authors should either justify the sudden-global-drop approximation quantitatively or estimate the effect of a finite transition width and mixed phase on Omega_a h^2.","section":"Eq. (2) and Fig. 2: homogeneous abrupt v_s drop"},{"comment":"The text states that 'the entire range f_a as 10^8-14 GeV is allowed', but Fig. 3 shows only a sparse set of colored circles and no table lists the corresponding Lambda, theta_i, and Omega_a h^2 values. Since Lambda is chosen per f_a to satisfy Eq. (14), the claim that the range is covered without fine-tuning theta_i needs at least a statement of the fixed theta_i used in the numerical runs and a denser scan or a continuous Lambda(f_a) curve. Without these, the reader cannot verify either the coverage of the quoted interval or the absence of fine-tuning in theta_i.","section":"Fig. 3 and claim of the full f_a range"}],"minor_comments":[{"comment":"The effective potential in Eq. (4) omits the Coleman-Weinberg and daisy contributions while retaining only the leading T^2 thermal terms; because T_n and v_s(T) directly set m_as(t) and the duration of stage II, a sensitivity check of the relic window to these corrections would materially strengthen the quantitative claims.","section":"Appendix B"},{"comment":"The value of theta_i used for the numerical solutions is not stated; the phrase 'without fine-tuning theta_i' should be quantified (for example, theta_i = 1 fixed, or varied only over an explicitly stated O(1) interval).","section":"Figs. 2 and 3"},{"comment":"The values of alpha_*, beta/H_*, and v_w used for the gravitational-wave spectra of BP1 and BP2 are not reported in the text; reporting these inputs would make the GW curves reproducible.","section":"Fig. 4 and Appendix C"},{"comment":"The number-density conservation step is applied from T_QCD onward, which is the standard approximation, but the text should explicitly note that the sudden switch of m_a(T) to the constant m_a0 at T_QCD in Eq. (2) is the point at which this conservation law is applied.","section":"Eq. (13)"}],"recommendation":"major_revision","confidential_remarks":"The decisive issue is the Eq. (1)/Eq. (11) inconsistency: the numerical results are for a different potential than the one defined by the Lagrangian. I did not attempt to reproduce the numerics, but the discrepancy is evident from the displayed equations and is not a small effect at theta_i ~ O(1). If the authors rerun the analysis with the correct period-pi potential and the broad f_a range survives, the paper would be a solid contribution; otherwise the current version cannot support its headline claim."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"First, the good news: the paper has a genuinely new idea. A first-order EWPT with an explicit PQ-breaking portal from the singlet S can give the QCD axion a temporary mass bump, then remove it, producing a second oscillation phase with a shifted field value. That recurrent misalignment mechanism is worth thinking about, and the paper is honest about its construction, with a UV completion and explicit benchmark points.\n\nThe bad news is the stress-test note is right, and it is prior to any other concern. The operator in Eq. (1) is S^4 Φ^2 + h.c., which for Φ ∝ e^{iθ} gives a potential ∝ cos(2θ+α). With α=π that is -cos2θ, period π. But the equation actually solved, Eq. (11), is θ¨ + 3Hθ˙ + m_a^2(T) sinθ = 0, the equation for a period-2π cosine. The paper adds the S contribution to the mass squared and puts it into that sinθ equation. That is not the dynamics of the stated operator. At θ=1 the force differs by nearly a factor of two, and for θ>π/2 the S potential has a minimum at π, which the sinθ equation does not. Every relic-satisfying point in Fig. 3 is computed with the wrong force. So the headline claim is not backed by the model as written.\n\nThe fix is straightforward: solve with the full potential, including the cos2θ term alongside the QCD potential. That is a numerical exercise, and the result may well preserve the qualitative story. But the numbers, the allowed f_a range, and the Λ values will shift. Until then, the central claim is unsubstantiated.\n\nSecondary soft spots are what the reader flagged: the homogeneous, instantaneous v_s drop is an idealization given bubbles, and Λ is fitted to relic rather than predicted. Those are addressable, but the equation mismatch is the load-bearing one.\n\nWho is this for? Axion dark matter phenomenologists and anyone working on EWPT/GW connections. The idea deserves a referee, but the paper needs major revision before it can be published. I would send it out, with a request that the authors redo the numerical analysis with the correct potential.","headline":"The recurrent misalignment idea is worth taking seriously, but the paper solves a different model from the one it defines: Eq. (1) gives a period-π axion potential, while Eq. (11) uses a period-2π sine, so the relic curves do not support the headline claim.","tokens_in":16458,"tokens_out":6849,"would_cite":false,"duration_ms":75541,"reading_group":"maybe","serious_thinker":"no","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"The QCD axion can account for all dark matter across $f_a\\in[10^8,10^{14}]$ GeV, without fine-tuning $\\theta_i$, if a first-order electroweak phase transition temporarily boosts its mass.","keywords":["QCD axion","dark matter","misalignment mechanism","recurrent misalignment","first-order electroweak phase transition","singlet scalar","Peccei-Quinn symmetry","gravitational waves"],"falsifier":"Recompute the axion relic density with a realistic nucleation history, including bubble fraction, percolation temperature, and a finite transition duration, replacing the instantaneous global drop of $v_s(T)$ at $T_n$; the mechanism is falsified if no $f_a$ in $[10^8,10^{14}]$ GeV then gives $\\Omega_a h^2=0.12$ with $\\theta_i\\sim 1$.","tokens_in":15264,"feed_emoji":"🌌","tokens_out":15435,"duration_ms":141804,"temperature":0.7,"pith_summary":"The paper's goal is to show that the dark matter window of the QCD axion can be dramatically widened if the electroweak phase transition is first order. The authors introduce a 'recurrent misalignment' mechanism in which a singlet scalar, present to make the transition first order, temporarily raises the axion mass through a non-renormalizable Peccei-Quinn-breaking operator. The axion consequently oscillates early, then stops as the singlet's vev drops, drifts through field space, and starts oscillating again with a shifted field value. The numerical solution of this two-stage evolution gives the observed relic abundance for $f_a\\in[10^8,10^{14}]$ GeV with an order-one misalignment angle, instead of the usual $10^{12}$ GeV point or fine-tuned $\\theta_i$. If correct, the result would motivate axion searches across a much wider mass range and link axion dark matter to gravitational-wave signals from the same phase transition.","feed_headline":"Early mass bump widens QCD axion dark-matter window to 1e8–1e14 GeV","feed_subtitle":"A temporary axion mass during a first-order electroweak transition lets θ≈1 match the observed relic density.","key_machinery":"The load-bearing object is the temporary mass term $$$m_a^{2}$(T)=m_{a0}^2(T)+\\frac{$4v_s^{4}$(T)}{\\$Lambda^{2}$}\\quad (T_n<T\\le T_s),$$ with $m_{a0}(T)$ the QCD instanton-generated axion mass and $v_s(T)$ the singlet vev from the high-temperature effective potential. The governing equation is the standard misalignment equation $\\ddot\\theta+3H\\dot\\theta+m_a^2(T)\\sin\\theta=0$, solved piecewise, with $T_s$, $T_c$ and $T_n$ fixed by the two-step singlet-assisted phase transition (benchmark nucleation temperatures are about 43 and 51 GeV). The abrupt disappearance of $v_s$ at $T_n$ is the pivotal event: it ends the early oscillation, leaves the axion with residual kinetic energy and a drifting field value, and thereby sets a new effective initial condition $\\theta_i^{II}$ for the second, QCD-driven oscillation. This two-stage oscillation sequence is the mechanism named recurrent misalignment.","core_discovery":"The paper's central claim is that a temporary, transition-induced axion mass can reset the initial condition for the late-time oscillation, so that the QCD axion saturates the dark matter relic for $f_a$ anywhere in $10^8$--$10^{14}$ GeV with $\\theta_i\\sim\\mathcal{O}(1)$. In the mechanism, a $Z_2$-symmetric real singlet $S$ gives the electroweak phase transition a tree-level barrier and makes it first order; while its vev $v_s(T)$ is nonzero between $T_s$ and the nucleation temperature $T_n$, a dimension-six operator $S^4\\Phi^2 e^{i\\alpha}/\\Lambda^2$ contributes $4v_s^4(T)/\\Lambda^2$ to the axion mass squared. The axion then begins a first oscillatory phase, stops when $v_s$ drops to zero at $T_n$, and, because it retains kinetic energy, continues moving in field space before the standard QCD mass restarts oscillations with a shifted field value $\\theta_i^{II}$ and a non-zero velocity. Choosing $\\alpha=\\pi$ keeps the CP-conserving minimum at $\\theta=0$, and the numerical integration of the axion equation of motion with this mass history yields $\\Omega_a h^2=0.12$ for a continuum of $(f_a,\\Lambda)$ pairs with $f_a<\\Lambda<M_{\\mathrm{Pl}}$.","pith_inferences":["The mechanism is not specific to the QCD axion: any axion-like particle with a temporary mass from a first-order transition should experience the same recurrent misalignment.","The abrupt global drop of $v_s(T)$ is an idealization; modeling the transition with finite bubble nucleation and percolation could shift the phase drift and should be checked before the full $10^8$--$10^{14}$ GeV range is taken as exact.","If a gravitational-wave signal from an electroweak-scale transition and an axion signal were both observed, the required $\\Lambda$ for each $f_a$ would provide a quantitative cross-check connecting the two observations.","The claimed upper limit $f_a\\le 10^{14}$ GeV is set by $\\Lambda<M_{\\mathrm{Pl}}$; relaxing the cutoff to $\\Lambda\\sim M_{\\mathrm{Pl}}$ or changing the transition strength would move this boundary, so the precise endpoint is model-dependent rather than fundamental."],"forward_implications":["The QCD axion can be all of the dark matter for $f_a$ between $10^8$ and $10^{14}$ GeV, corresponding to axion masses from tens of neV to tens of meV, with $\\theta_i$ of order one and no fine-tuning.","Each allowed $f_a$ comes with a specific cutoff scale $\\Lambda$ satisfying $f_a<\\Lambda<M_{\\mathrm{Pl}}$; measuring the axion mass would thus pin down the scale of the PQ-breaking portal.","The same first-order transition that drives the mechanism generates a stochastic gravitational-wave background in the mHz-to-Hz band, so the axion dark-matter prediction is tied to a foreground observable by next-generation gravitational-wave detectors.","Because $\\alpha=\\pi$ keeps the minimum CP-conserving, the mechanism does not re-introduce the strong CP problem after $v_s$ relaxes to zero.","The widened mass range places relic-satisfying QCD axions within the reach of a broader set of haloscope-style detection schemes than the standard $\\sim\\mu$eV window."],"supporting_citations":[{"why":"It supplies the singlet-extended scalar potential whose tree-level barrier makes the electroweak phase transition first order.","marker":"[40]"},{"why":"It supplies the temperature-dependent QCD axion mass behavior used before, during, and after the transition.","marker":"[51]"},{"why":"It provides the bounce-action computation that fixes the nucleation temperature $T_n$ for the benchmark points.","marker":"[56]"},{"why":"It furnishes the coherent-oscillation equation and relic-density formalism that the paper integrates.","marker":"[10]"},{"why":"It is the authors' earlier treatment of a similar mass-changing operator, whose crossover limitation motivates the singlet-based first-order transition.","marker":"[41]"},{"why":"It provides the supernova bound $f_a\\gtrsim 10^8$ GeV that fixes the lower end of the claimed dark-matter range.","marker":"[12]"},{"why":"It supplies the experimental sensitivity curves against which the relic-satisfying axion points are displayed.","marker":"[57]"}],"fun_headline_variants":["Temporary mass bump during first-order EWPT frees axion dark matter","First-order EWPT sets axion angle naturally, no fine-tuning","Gravitational waves reveal axion dark matter via first-order EWPT","Axion dark matter window from 1e8 to 1e14 GeV without fine-tuning"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The argument assumes the singlet's vev behaves as a single spatially uniform background that switches off abruptly at the nucleation temperature, although a real first-order transition proceeds through coexisting bubbles and takes finite time.","fun_headline_variants_meta":{"raw":{"variants":["Temporary mass bump during first-order EWPT frees axion dark matter","First-order EWPT sets axion angle naturally, no fine-tuning","Gravitational waves reveal axion dark matter via first-order EWPT","Axion dark matter window from 1e8 to 1e14 GeV without fine-tuning"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00108,"raw_usage":{"total_tokens":4560,"prompt_tokens":1030,"completion_tokens":3530,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":646,"completion_tokens_details":{"reasoning_tokens":3446}},"tokens_in":646,"tokens_out":3530,"duration_ms":25420,"temperature":1.0,"reasoning_tokens":3446,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T19:29:33.915310+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Recompute the axion relic density with a realistic nucleation history, including bubble fraction, percolation temperature, and a finite transition duration, replacing the instantaneous global drop of $v_s(T)$ at $T_n$; the mechanism is falsified if no $f_a$ in $[10^8,10^{14}]$ GeV then gives $\\Omega_a h^2=0.12$ with $\\theta_i\\sim 1$.","supporting_citations":[{"cited_title":"Effects of Electroweak Symmetry Breaking on Axion Like Particles as Dark Matter","cited_arxiv_id":"2311.05125","evidence_quote":"It is the authors' earlier treatment of a similar mass-changing operator, whose crossover limitation motivates the singlet-based first-order transition."}],"review_version":1}