{"id":"2182a6d6-3244-43d4-ad2d-2e591991e5d2","arxiv_id":"2506.19918","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":8,"one_line_summary":"A supercooled first-order phase transition near the QCD scale can briefly set the QCD axion into mini kinetic misalignment and, through bubble inhomogeneities, produce axion domain walls even without cosmic strings.","lead":"This paper studies what happens to the QCD axion, a dark matter candidate, when a supercooled first-order phase transition reheats the universe near the QCD scale. It predicts a brief axion rotation phase and, in part of parameter space, axion domain walls that would overclose the universe.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Eq. (9) contains a sign/power error in the trapping-temperature factor: since K/V grows as T^14, Ttrap/TLambda = (K/V)^{-1/14}, not (K/V)^{1/14}, which can shift the DW boundaries by large factors.","rationale":"The paper is a legitimate exploratory phenomenological study: it identifies a new way a supercooled FOPT near the QCD scale can change axion dynamics and produce domain walls without cosmic strings. The EOMs are standard, the 1D simulation is a useful first step, and the qualitative direction of the mechanism is plausible. However, the central quantitative claim rests on Eq. (9), and that equation contains a concrete power-counting error in the Ttrap/TLambda factor: because K/V grows as T^14, the trapping ratio is the inverse of what is printed. This is not a matter of outside consensus; it is an internal dimensional/power-check inconsistency. If the numerical Fig. 6 was actually generated with the inverse factor, the error is only typographical, but the paper should say so explicitly. If it was generated with the printed factor, the claimed DW parameter space is overestimated. The reader's CONDITIONAL verdict remains appropriate: the mechanism is not rejected, but its most load-bearing quantitative condition needs a direct check before the overclosure constraint can be trusted.","tokens_in":9728,"tokens_out":23068,"duration_ms":276250,"concrete_test":"Recompute Fig. 6 using Ttrap/TLambda = (K(TLambda)/Vmax(TLambda))^{-1/14} instead of the printed +1/14 power, keeping all other inputs (theta_a,i, beta/HLambda, TPT, fa, HLambda grids) identical. If the DW and extended-DW boundaries shift by more than about 10% in either axis, Eq. (9) and the overclosure region need revision. As a cheaper analytic control, integrate the homogeneous axion EOM (5) from TLambda to Ttrap for several Fig. 6 benchmarks and compare the actual angle at trapping with the product in Eq. (9).","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central domain-wall condition Eq. (9) is not yet quantitatively secure because the redshift factor Ttrap/TLambda in it is written with the wrong power. The text states Ttrap/TLambda = (K(TLambda)/Vmax(TLambda))^{1/14}. But with K proportional to a^{-6} and Vmax proportional to m_a(T)^2 proportional to T^{-8} proportional to a^8 (for T > Lambda_QCD), the ratio K/Vmax is proportional to a^{-14} proportional to T^{14}. Since Eq. (4) and Fig. 6 impose K(TLambda)/Vmax(TLambda) > 1, the trapping temperature lies below TLambda, so the correct relation is Ttrap/TLambda = (K(TLambda)/Vmax(TLambda))^{-1/14}. Using the printed inverse factor can change the amplitude estimate in Eq. (9) by orders of magnitude (for example, a factor of about 100 when K/V ~ 10^14), directly moving the DW-formation and extended-DW boundaries in Fig. 6. In addition, Eq. (9) treats (Ttrap/TLambda) as a simple redshift of a free massless amplitude, but the phase at trapping is set by the nonlinear trapping dynamics; the only numerical check is a single 1D multi-bubble run (Fig. 4), and the paper explicitly defers full 3D merging dynamics. Thus the quantitative borders of the central claim are not yet established.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper studies the QCD axion cosmology in a universe where a supercooled first-order phase transition (FOPT) in a dark scalar sector reheats the Standard Model bath to temperatures near or above the QCD scale. Under pre-inflationary PQ breaking, all Hubble patches initially share a common axion angle. The authors propose that the reheating event temporarily flattens the axion potential, inducing a transient 'mini kinetic misalignment' phase in which the axion field rotates. They then argue that stochastic bubble nucleation produces spatial variation in the reheating history, so different bubbles develop different axion phases; when the axion potential regrows, domain walls form even in the absence of cosmic strings. The paper derives an order-of-magnitude criterion, Eq. (9), supports it with a one-dimensional multi-bubble simulation, and maps the DW-forming regions on the H_Lambda-f_a plane for two reheating scenarios in Fig. 6. It concludes that infinitely extended domain walls in the feeble-coupling case enter scaling and eventually overclose the universe, while finite enclosed walls collapse into axion radiation.","tokens_in":10180,"tokens_out":7627,"duration_ms":82606,"significance":"If the mechanism works, it is significant: it connects PTA/NANOGrav-motivated supercooled FOPTs to axion cosmology, produces a new channel for QCD axion domain walls without cosmic strings, and could impose strong constraints on the FOPT parameter space. The paper has real strengths: the homogeneous WKB evolution is standard, the 1D multi-bubble simulation directly illustrates the proposed effect, and the authors are explicit about the assumptions that go into the domain-wall criterion. At the same time, the quantitative boundaries of the central claim rest on an order-of-magnitude formula and on 3D merging dynamics that are not simulated, so the manuscript currently establishes the qualitative existence of the effect more firmly than it establishes the precise parameter-space regions.","major_comments":[{"comment":"The trapping-temperature ratio is stated with the wrong power of the kinetic-to-potential ratio. With K proportional to a^{-6} and Vmax proportional to m_a(T)^2 proportional to T^{-8} proportional to a^8 after reheating, K/Vmax is proportional to T^{14}. Since the mini-kinetic-misalignment condition in Eq. (4) imposes K(T_Lambda)/Vmax(T_Lambda) > 1, the trapping temperature lies below T_Lambda, so the correct relation is Ttrap/T_Lambda = (K(T_Lambda)/Vmax(T_Lambda))^{-1/14}, not the printed (K(T_Lambda)/V(T_Lambda))^{1/14}. The positive power overestimates the redshift factor by a factor of (K/V)^{2/14}, which is about 10^2 for K/V ~ 10^{14}; through Eq. (9) this directly shifts the DW-formation and extended-DW boundaries in Fig. 6 by orders of magnitude. This needs to be corrected and Fig. 6 redrawn before the quantitative claims can be assessed.","section":"Section 'Axion Field Evolution - Homogenous Background', text after Eq. (4); Eq. (9)"},{"comment":"The central claim of domain-wall formation, and especially of infinitely extended walls, is supported by a single 1D multi-bubble simulation, while the paper explicitly states that 'Modeling the full three-dimensional dynamics of bubble merging is highly nontrivial and lies beyond the scope of this work.' The extended-infinite-DW region in Fig. 6 is obtained by applying a 3D percolation threshold (p_c = 0.31) to randomly assigned phase signs, but the 1D simulation cannot validate that this sign distribution survives 3D bubble collisions, wall curvature, and gradient-energy draining. If 3D effects suppress the coherent rotation or correlate the phases, the extended-DW region would shrink or disappear. Because this region is the basis of the overclosure constraint, this is a load-bearing gap rather than a presentation issue.","section":"Section 'Axion Field Evolution - Inhomogeneity from Bubbles'; Fig. 4"},{"comment":"Eq. (9) is an order-of-magnitude criterion with an O(1) threshold (pi), and the phase at trapping is set by nonlinear dynamics rather than by the simple free-massless redshift factor used in the formula. The paper tests this criterion with only one 1D benchmark (Figs. 3-4) and does not scan over the axion phase phi_PT, the bubble radius r_s, or the transition strength beta/H_Lambda. Nevertheless, Fig. 6 draws sharp boundaries for the DW-formation and extended-DW regions without any uncertainty band or sensitivity estimate. The figure should be replaced or supplemented by a band or multi-benchmark scan so that the claimed exclusion regions are not read as precise predictions.","section":"Fig. 6 and Eq. (9)"}],"minor_comments":[{"comment":"Equation (1) is typeset in a way that is very difficult to parse; the relation between the two scenario labels, the axion mass ratio, and the condition m_a(Tosc) ~ 3Hosc should be written out explicitly.","section":"Eq. (1)"},{"comment":"The notation 'V(T_Lambda)' should be 'Vmax(T_Lambda)' for consistency with the definition given just before Eq. (4).","section":"Text after Eq. (4)"},{"comment":"Figure 2 has no colorbar or explicit scale for the filled contours, so the reader cannot quantitatively read the reset theta_a values from the plot.","section":"Fig. 2"},{"comment":"The step-function expression for m_a uses r_s + delta_t, but the time variable delta_t is not defined clearly; the text should state explicitly that delta_t is the time elapsed since reheating of the bubble interior.","section":"Section 'Axion Field Evolution - Inhomogeneity from Bubbles', around Eq. (8)"},{"comment":"There are a few language and typographical issues, such as 'the hight of the potential energy barrier' and 'in the aftermentioned maintext'; these should be corrected in a revised version.","section":"Throughout"}],"recommendation":"major_revision","confidential_remarks":"The core idea is novel and worth pursuing, and the paper is refreshingly honest about its limitations. However, the sign error in the Ttrap/T_Lambda factor directly changes Fig. 6, and the 3D merging dynamics are essential to the extended-DW statement. I would encourage a revision that corrects Eq. (9), redraws Fig. 6, and adds a sensitivity scan over the free parameters; these are feasible within the scope of the manuscript."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nThe one thing to know: this is a genuinely new mechanism — QCD axion domain walls produced by the inhomogeneous reheating from a supercooled FOPT — but the paper's central condition, Eq. (9), has a redshift-factor error that could shift the claimed DW boundaries by orders of magnitude. The qualitative idea is worth taking seriously; the quantitative map in Fig. 6 is not yet reliable.\n\nWhat's new and good: The authors correctly identify that in a pre-inflation PQ-breaking scenario, a supercooled FOPT reheating the SM above Λ_QCD first sends the axion into a transient 'mini kinetic misalignment' phase, and then the stochastic spread in bubble reheating times creates spatial gradients in the axion phase. When the axion mass turns on, those gradients can freeze into domain walls in the absence of cosmic strings. That wall-from-bubble-fluctuation mechanism is not in the cited kinetic misalignment or FOPT papers. The homogeneous WKB evolution is standard and correctly applied; the 1D multi-bubble simulation in Fig. 4 is a reasonable sanity check and shows the claimed pattern. The paper is honest about leaving 3D merger dynamics to future work.\n\nSoft spots, in order of severity. First, the sign/power error: the text states T_trap/T_Λ = (K(T_Λ)/V(T_Λ))^{1/14}. But with K ∝ a^{-6} and V ∝ m_a^2 ∝ a^8, the ratio K/V ∝ a^{-14} ∝ T^{14}. Since K/V > 1 at T_Λ, the correct relation is T_trap/T_Λ = (K/V)^{-1/14}. The printed version puts the trapping temperature above T_Λ and the amplitude in Eq. (9) too large; for K/V ~ 10^14 the amplitude is off by a factor ~100. That directly moves the 'DW formation' and 'extended DW' boundaries in Fig. 6. Second, Eq. (9) is an order-of-magnitude estimate that treats the trapped amplitude as a simple redshifted free-field value, but the phase at trapping is set by nonlinear dynamics; the 1D simulation is supportive but not a substitute for a proper 3D treatment. Third, the benchmark choices (θ_a,i = 2, T_PT = T_QCD or T_Λ/2) are illustrative; the boundaries carry no uncertainties. None of this invalidates the qualitative mechanism, and there is no circular reasoning — the condition is derived from assumed parameters, not fitted to the target outcome.\n\nWho this is for: anyone working on axion cosmology in nonstandard thermal histories or interpreting the PTA signal with supercooled FOPTs. The paper deserves a serious referee — the new mechanism and its cosmological implications are important enough — but the referee should require the redshift correction and a 3D consistency check before the quantitative claims stand. I would not cite Fig. 6 as it currently stands.","headline":"New mechanism for QCD axion domain walls from FOPT reheating inhomogeneity, but Eq. (9) has a sign/power error in the trapping-temperature redshift factor that shifts the claimed boundaries by orders of magnitude; worth refereeing, not yet quantitatively reliable.","tokens_in":10610,"tokens_out":3021,"would_cite":false,"duration_ms":28256,"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":"A supercooled first-order phase transition near the QCD scale can send the QCD axion through a brief kinetic rotation and then, via stochastic bubble reheating, create axion domain walls even without cosmic strings.","keywords":["QCD axion","domain walls","first-order phase transition","kinetic misalignment","axion dark matter","pulsar timing array","bubble nucleation","overclosure"],"falsifier":"A full three-dimensional lattice simulation of stochastic bubble nucleation, merger, and axion field evolution would settle the claim: if the maximum axion amplitude reached in any region before trapping stays below $\\pi$ for parameter points inside the dark region of Fig. 6, the domain-wall condition (Eq. (9)) fails and the overclosure conclusion does not follow.","tokens_in":9548,"feed_emoji":"🌌","tokens_out":17891,"duration_ms":152589,"temperature":0.7,"pith_summary":"The paper claims that a supercooled first-order phase transition near the QCD scale—the kind that could explain the recently reported pulsar-timing-array gravitational-wave excess—significantly alters the early-universe behavior of the QCD axion. When the transition reheats the Standard Model bath, the axion potential suddenly flattens, so the field's existing velocity sends it rotating through its potential in a short 'mini kinetic misalignment' stage. Because bubble nucleation is stochastic, different bubbles reheat at slightly different times, leaving different axion phases in different regions; when the QCD potential reasserts itself, the axion settles into different minima in neighboring regions, forming domain walls. These walls appear even though no cosmic strings exist, because the axion's global symmetry was broken before inflation. In the parameter range where the walls are infinitely extended, they enter a scaling regime and overclose the universe, so that combination of axion and phase transition is observationally excluded.","feed_headline":"Axion domain walls form from supercooled bubble reheating","feed_subtitle":"The same reheating that explains pulsar-timing signals may make axion domain walls overclose the universe.","key_machinery":"The mechanism is the coupling between the sharply temperature-dependent axion potential and the spatially stochastic reheating of bubble nucleation. Before the transition the axion mass grows as $m_a(T)\\propto T^{-4}$, so reheating above $\\Lambda_{\\rm QCD}$ flattens the potential and turns the field's oscillation into a rotation; this is the 'mini kinetic misalignment' stage. Bubble nucleation is stochastic, so each bubble reheats at a slightly different time and the axion velocity phase at reheating differs from bubble to bubble. Continuity pins the field at each bubble wall near $\\theta_a=0$, and the interior rotation sends a gradient wave inward; condition (9), $m_a(T_{\\rm PT})\\langle\\theta_a\\rangle/\\beta\\,(T_{\\rm trap}/T_\\Lambda)>\\pi$, states when the inward-propagating amplitude before trapping exceeds half a period, seeding different minima in different regions. The percolation threshold $p_c=0.31$ then determines whether the resulting walls are finite closed bubbles or infinitely extended walls that enter the scaling regime.","core_discovery":"The paper's central claim is that spontaneous breaking of the axion's global symmetry before inflation, followed by a supercooled first-order phase transition that reheats the Standard Model above the QCD confinement scale, converts the conventional axion misalignment picture into a rotation-dominated stage. The axion's temperature-dependent mass, $m_a(T)\\propto T^{-4}$, makes the potential flatten sharply upon reheating, so the field's velocity at the phase-transition moment can carry it over potential barriers in the mini kinetic misalignment stage. Because bubble nucleation is stochastic, each bubble interior reheats at a slightly different time, so the axion's velocity phase differs from bubble to bubble; the bubble wall, pinned near $\\theta_a=0$ by continuity, launches inward-propagating gradients. If the amplitude accumulated before trapping satisfies $m_a(T_{\\rm PT})\\langle\\theta_a\\rangle/\\beta\\,(T_{\\rm trap}/T_\\Lambda)>\\pi$ (Eq. (9)), the axion ends up in different vacuum values in different regions, producing QCD axion domain walls decoupled from cosmic strings. In the feeble reheating case, these walls can be infinitely extended when the percolation threshold is crossed; their tension, $\\sim m_{a,0}f_a^2$, far exceeds the observational upper bound, so they dominate the universe and contradict observation.","pith_inferences":["We speculate that the same boundary-condition logic applies to any axion-like particle with a steeply temperature-dependent mass, so the $H_\\Lambda$-$f_a$ plane in Fig. 6 could serve as a template for broader ALP cosmologies, though the paper only treats the QCD axion.","We suspect the 1D simulation may exaggerate the coherence of the rotating interior; a full 3D nucleation-merger simulation could shrink the domain-wall region in Fig. 6 or erase it entirely if gradient energy escapes before trapping.","If finite closed walls do form and collapse, their decay into axions and gravitational waves at a specific epoch would be an observable signature that connects the phase transition to axion physics; this is an extension the paper mentions as future work but does not itself establish."],"forward_implications":["The axion relic abundance can no longer be read off the initial misalignment angle alone: the phase transition resets the effective misalignment angle, shifting the standard relation between $\\theta_{a,i}$ and the dark-matter density.","Domain walls form without cosmic strings in a sizeable part of the $H_\\Lambda$-$f_a$ plane shown in Fig. 6, undermining the usual assumption that pre-inflation symmetry breaking guarantees freedom from axion domain walls.","In the regions where the walls are infinitely extended, they enter the scaling regime and overclose the universe, so the pulsar-timing-motivated supercooled phase transition is excluded there in the presence of the QCD axion.","In the feeble-coupling case, extended infinite walls form when the volume fraction of the selected vacuum exceeds the percolation threshold; in the instantaneous-reheating case, only small enclosed walls with size $\\sim m_{a,0}^{-1}$ can form, so the overclosure bound does not apply in that limit."],"supporting_citations":[{"why":"It establishes the stochastic bubble nucleation picture that produces bubble-to-bubble reheating variation.","marker":"[29, 30]"},{"why":"It gives the WKB conservation law used to set the axion field amplitude and velocity at the phase transition.","marker":"[31]"},{"why":"It introduces kinetic misalignment, the foundation for the mini kinetic misalignment stage.","marker":"[32–34]"},{"why":"It provides the bubble formation-time distribution peaked at $\\beta^{-1}$, setting the timescale in the wall-formation condition.","marker":"[35]"},{"why":"It classifies domain walls as finite enclosed or infinitely extended, defining the two outcomes.","marker":"[36]"},{"why":"It supplies the scaling-regime evolution used to conclude infinite walls overclose the universe.","marker":"[37, 38]"},{"why":"It gives the observational bound on domain-wall tension that makes the axion walls inconsistent with observation.","marker":"[39]"},{"why":"It provides the percolation transition threshold $p_c=0.31$ used to map where infinite walls appear.","marker":"[40]"}],"fun_headline_variants":["Supercooled bubbles make axion walls that dominate","Axion walls and pulsar timing hints may conflict","Mini kinetic misalignment yields axion domain walls","Bubble reheating can trap axions into domain walls","Supercooled phase transition links NANOGrav and axion walls"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The argument depends on each reheated bubble interior behaving as a coherently rotating axion patch whose boundary at the bubble wall stays pinned near zero field, and on that coherent rotation surviving the full three-dimensional merger of many bubbles; the paper checks this only with a one-dimensional simulation.","fun_headline_variants_meta":{"raw":{"variants":["Supercooled bubbles make axion walls that dominate","Axion walls and pulsar timing hints may conflict","Mini kinetic misalignment yields axion domain walls","Bubble reheating can trap axions into domain walls","Supercooled phase transition links NANOGrav and axion walls"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000274,"raw_usage":{"total_tokens":1657,"prompt_tokens":981,"completion_tokens":676,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":597,"completion_tokens_details":{"reasoning_tokens":596}},"tokens_in":597,"tokens_out":676,"duration_ms":6551,"temperature":1.0,"reasoning_tokens":596,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T18:24:44.075289+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A full three-dimensional lattice simulation of stochastic bubble nucleation, merger, and axion field evolution would settle the claim: if the maximum axion amplitude reached in any region before trapping stays below $\\pi$ for parameter points inside the dark region of Fig. 6, the domain-wall condition (Eq. (9)) fails and the overclosure conclusion does not follow.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"It gives the observational bound on domain-wall tension that makes the axion walls inconsistent with observation."}],"review_version":2}