{"id":"86524cc8-f149-467d-a939-4863d97f0cbf","arxiv_id":"2501.15236","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"high","formal_verification":"none","parameter_count":3,"one_line_summary":"A new Mathematica package solves the first-order phase transition bounce equation with a field-dependent wavefunction renormalization factor, showing that the factor can substantially modify both the bounce action and the tunneling path.","lead":"The authors release a Mathematica package, VacuumTunneling, that computes the Euclidean bounce action for first-order phase transitions while including the wavefunction renormalization factor Z, which earlier bounce packages omit. The package also handles numerical potentials and supercooled thick-wall transitions, and the paper reports that including Z changes the action and the tunneling path.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Eq. (2.34) is dimensionally inconsistent: the NLO perpendicular force uses (d^2x/dr^2)∇⊥logZ instead of (dx/dr)^2∇⊥logZ, so the multi-field path-deformation algorithm does not solve the stated NLO equation of motion.","rationale":"The most load-bearing condition for the central claim is not Z>0 alone. The authors explicitly state the Z<0/zero-crossing limitation for EWPT and defer that to future work, so that is a known restriction, not an internal inconsistency. The decisive problem is the multi-field NLO equation of motion and its perpendicular projection: (2.27) drops a term that vanishes only in the single-field limit, and (2.34) has a dimensionally wrong coefficient. This directly undermines the package's advertised extension to multi-field tunneling with renormalization. The single-field LO validation (agreement with CosmoTransitions, FindBounce, AnyBubble) and the plausible single-field NLO implementation provide independent evidence that part of the code is sound, but they cannot certify the multi-field NLO path deformation. The reader's CONDITIONAL verdict already asks for correction of the perpendicular-force derivation and verification of the code; my concern is a sharper version of the same condition, so I do not change the verdict. I set agreement_with_reader to partial because the reader's formal weakest assumption was the positivity of Z, whereas the more load-bearing flaw is the multi-field equation error.","tokens_in":24267,"tokens_out":21965,"duration_ms":181186,"concrete_test":"Re-derive the perpendicular projection from the Euler–Lagrange equations of (2.5), or inspect the package source around the calculation of N, to determine whether the code multiplies ∇⊥logZ by (dx/dr)² or by d²x/dr². If the code follows (2.35), rerun the two-field example in §IV.B with N corrected to (dx/dr)² d²σ/dx² + (1/2)(dx/dr)²∇⊥logZ − Z∇⊥V and compare the action with the printed 71283.7; a material shift confirms the printed NLO multi-field results are not solving the stated equation. If the code already uses (dx/dr)², the fix is textual and the existing numerical results can be kept.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Varying (2.5) for multiple fields gives σ'' + (D−1)r^{-1}σ' − (∇logZ·σ')σ' + (1/2)∇logZ|σ'|² = Z∇V. Eq. (2.27) omits the projection term (∇logZ·σ')σ'; in one dimension the missing term combines with the existing one to reproduce (2.6), so single-field tests cannot detect the error. For a path parameterized by arc length, σ' = T dx/dr and σ'' = T d²x/dr² + T'(dx/dr)². The perpendicular component of the correct EOM is (dx/dr)² T' + (1/2)(dx/dr)²∇⊥logZ = Z∇⊥V. The paper's Eq. (2.34) instead has −(1/2)(d²x/dr²)∇⊥logZ on the left. The latter term has dimension 1/L² while all other terms have dimension [field]/L², so (2.34) cannot be the perpendicular projection. Hence N in (2.35) and the path deformation (2.38) do not enforce (2.27) for Z≠1. The NLO multi-field actions in Table I and Fig. 5 therefore do not support the central claim unless the code uses a corrected formula that the text does not provide.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper presents a Mathematica package, VacuumTunneling, for computing Euclidean bounce actions in first-order phase transitions with a field-dependent wavefunction renormalization factor Z in the derivative-expanded effective action. The single-field solver is a modified shooting method that includes the extra velocity-squared term in Eq. (2.6); the multi-field solver is a path-deformation method that adjusts a path in field space until a 'perpendicular force' N in Eq. (2.35) vanishes. The paper validates the Z=1 limit against CosmoTransitions, FindBounce, and AnyBubble (action agreement within 1%), and applies the package to a two-field model, random multi-field potentials with 2-8 fields, a supercooled conformal model, and a PNJL chiral phase transition. The paper also documents optimizations for supercooled transitions and for numerically defined potentials and renormalizations.","tokens_in":24613,"tokens_out":15151,"duration_ms":133334,"significance":"The package fills a genuine gap: existing public codes solve the Z=1 bounce equation, whereas NLO effective actions contain a field-dependent wavefunction renormalization. The single-field NLO equation (2.6) is plausibly derived, and the Z=1 validation against three independent packages is real external support; the supercooling optimization and support for numerical input are useful features. The paper is also explicit about the restriction to positive Z and about the need for regularization in electroweak applications. However, the multi-field Z-neq-1 results are not currently supported, because the perpendicular projection equation (2.34) is dimensionally inconsistent and the multi-field equation of motion (2.27) is not the variation of (2.5). Since the headline claim that the renormalization factor changes the tunneling path rests on this multi-field algorithm, the central advertised capability is unverified as written.","major_comments":[{"comment":"The multi-field equation of motion stated in Eq. (2.27) is not the Euler-Lagrange equation of the action in Eq. (2.5). Varying S = integral r^{D-1}[Z^{-1}|sigma'|^2/2 + V] gives sigma'' + (D-1)/r sigma' - (nabla log Z dot sigma') sigma' + (1/2) nabla log Z |sigma'|^2 = Z nabla V. Equation (2.27) instead contains only the term -(1/2) nabla log Z |sigma'|^2 and omits the projection term -(nabla log Z dot sigma') sigma'. For a single field the two terms combine into -(1/2) partial_logZ/partial_sigma sigma'^2 and Eq. (2.6) is recovered, which is why single-field tests cannot detect the problem. In multi-field tunneling the omitted term is generally nonzero, so the NLO multi-field paths and actions reported in Table I and Fig. 5 are not computed from the stated action.","section":"II.C, Eq. (2.27)"},{"comment":"The perpendicular projection is dimensionally inconsistent. Writing sigma(r) = sigma(x(r)) with |d sigma/dx| = 1 gives sigma' = T dx/dr and sigma'' = T d^2x/dr^2 + (dT/dx)(dx/dr)^2, with T = d sigma/dx. Projecting the correct equation perpendicular to T yields (dx/dr)^2 d^2 sigma/dx^2 + (1/2)(dx/dr)^2 nabla_perp log Z = Z nabla_perp V. Equation (2.34) instead has -(1/2)(d^2x/dr^2) nabla_perp log Z; this term has dimension 1/(L^2 [field]) while all other terms have dimension [field]/L^2. Consequently the perpendicular force N in Eq. (2.35) and the path deformation (2.38) do not enforce Eq. (2.27) when Z is not equal to 1. The Z=1 validation against FindBounce cannot detect this because the offending term vanishes for constant Z.","section":"II.C, Eq. (2.34)"},{"comment":"The multi-field Z-neq-1 numbers in Table I and the modified path in Fig. 5 are the paper's main evidence for the claim that Z significantly changes the tunneling path and action. Because the algorithm's dynamical equations are the ones criticized above, these numbers cannot currently be interpreted as solutions of the stated NLO problem. A revision should correct Eqs. (2.27) and (2.34), re-run the multi-field examples, and add an independent check of the corrected multi-field solver, for example by comparing a two-field Z-neq-1 solution against a direct numerical solution of the full boundary-value problem.","section":"Table I and Fig. 5"}],"minor_comments":[{"comment":"The abstract contains a typo: 'center rule' should be 'central role', and 'as well on the tunneling path' should be 'as well as on the tunneling path'.","section":"Abstract"},{"comment":"The heading of Section IV.B reads 'tow-field tunneling' and should be 'two-field tunneling'; the caption of Fig. 5 contains the stray text 'renormalizat]on'.","section":"IV.B and Fig. 5"},{"comment":"In the right panel of Fig. 4 the table header reads 'packge action'; it should be 'package action'.","section":"IV.A, Fig. 4"},{"comment":"The sentence 'we would like to redefine the renormalization factor Z by 1/Z' is ambiguous; clarify whether Z_sigma in Eqs. (2.5)-(2.6) is the inverse of the function called Z in Eq. (2.1), since the same symbol is used for both.","section":"II.B, Eq. (2.5)"},{"comment":"The Introduction states that the package works 'at any dimension', while the Program Summary restricts the implementation to D = 3, 4; please reconcile these statements.","section":"I and Program Summary"}],"recommendation":"major_revision","confidential_remarks":"The advertised multi-field NLO capability is the part that most distinguishes this package from existing software, and it is the part that fails verification. The single-field NLO part appears sound and the Z=1 validation is genuine. The revision should focus on correcting the equations of motion and the perpendicular force, and on validating the corrected Z-neq-1 multi-field solver against an independent method in a two-field example."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick take: the first public bounce solver to include a field-dependent kinetic renormalization Z; the single-field part is solid and LO-validated, but the multi-field NLO path-deformation equation has a dimensional inconsistency that must be fixed before those results can be trusted.\n\nWhat is actually new: the package itself. Existing tools (CosmoTransitions, BubbleProfiler, AnyBubble, FindBounce, SimpleBounce) all solve the Z=1 equation. This one adds a shooting method for the modified single-field equation with the -(1/2)(d log Z/dφ)(dφ/dr)^2 term and a path-deformation extension for multiple fields. The single-field Z=1 mode is checked against three independent packages and agrees within 1% on the action; that is genuine external support. The authors also state their limits plainly: Z must stay positive between the vacua, and for electroweak transitions gauge fields make Z cross zero, which introduces a singularity they do not regularize.\n\nThe main problem is the multi-field NLO derivation. Projecting the correct multi-field EOM, Eq. (2.27), perpendicular to the path gives -(1/2)(dx/dr)^2 ∇⊥ log Z, not -(1/2)(d^2 x/dr^2) ∇⊥ log Z as printed in Eqs. (2.34) and (2.35). The printed term has different dimensions from the rest of the equation, so the perpendicular force N in Eq. (2.35) does not enforce the stated EOM when Z differs from 1. Single-field tests cannot catch this because in one dimension the tangential and perpendicular projections recombine into the correct scalar equation. As a result, the multi-field NLO actions in Table I, the path shifts in Fig. 5, and the path-deformation result in the chiral phase transition example are not established by the text as written. If the code actually uses the corrected formula, the fix is textual; if it uses the printed formula, the multi-field NLO results are wrong. Either way, the authors should verify the code against an independent derivation and add convergence benchmarks (e.g., action versus PointsNumber), which are currently missing.\n\nMinor soft spots: the NLO multi-field examples have no error estimates, and the LO validation is only at the 1% level. The supercooling optimization and the two physical applications are useful extras, but they do not substitute for validation of the central NLO claim.\n\nWho this is for: anyone computing cosmological first-order phase transition rates who needs a bounce action with a renormalization factor. It deserves a serious referee: the LO part is well grounded, the flaw is specific and fixable, and the package fills a real gap. A referee should ask for the corrected derivation, an independent cross-check on a simple two-field test, and convergence data.","headline":"First public bounce solver with field-dependent Z; single-field LO-validated, but multi-field NLO path-deformation equation has a dimensional inconsistency that must be fixed.","tokens_in":25108,"tokens_out":8911,"would_cite":true,"duration_ms":72525,"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 paper presents a Mathematica package, VacuumTunneling, that computes the Euclidean bounce action for first-order phase transitions with the wavefunction renormalization factor $Z$ included, and demonstrates that $Z$ can significantly…","keywords":["vacuum decay rate","bounce action","wavefunction renormalization factor","first-order phase transition","shooting method","path deformation","supercooling phase transition","Mathematica package"],"falsifier":"Directly minimize the next-to-leading-order action functional $S=\\int dr\\,r^{D-1}\\left(\\frac{Z^{-1}}{2}\\sigma'^2 + V_{\\rm eff}\\right)$ for a single-field model with a non-constant $Z$ using an independent global method, such as fine-grid relaxation or an analytic solution where one exists, and compare the action and profile with the package's output; agreement to the claimed one-percent accuracy would confirm that Eq. (2.6) is being solved, while a larger mismatch would falsify the solver's central claim.","tokens_in":24056,"feed_emoji":"🌌","tokens_out":12546,"duration_ms":100822,"temperature":0.7,"pith_summary":"The paper claims that computing vacuum decay rates at next-to-leading order requires solving the bounce equation that comes from the derivative-expanded effective action with the wavefunction renormalization factor $Z$ retained, not just the leading-order equation with $Z=1$. To make this possible, the authors build a Mathematica package, VacuumTunneling, which solves the modified single-field bounce equation by shooting and extends it to multi-field tunneling by path deformation, including the extra friction-like term $-\\frac{1}{2}\\frac{d\\log Z}{d\\sigma}\\left(\\frac{d\\sigma}{dr}\\right)^2$. They show that the presence of $Z$ changes both the tunneling path in field space and the value of the bounce action; in one two-field example the action changes by more than two orders of magnitude, while at $Z=1$ the package reproduces the actions of existing packages within about one percent. A sympathetic reader would care because a consistent next-to-leading-order treatment is needed to address the gauge dependence of tunneling rates, and because some phase transitions have kinetic terms that vanish at leading order, making the $Z$ factor the leading contribution to the bounce.","feed_headline":"Renormalization factor changes vacuum tunneling action and path","feed_subtitle":"A new package solves next-to-leading-order bounce equations, showing Z alters action, path, and rates.","key_machinery":"The load-bearing object is the modified Euclidean bounce equation for an $O(D)$-symmetric bubble, Eq. (2.6): $\\frac{d^2\\sigma}{dr^2}+\\frac{D-1}{r}\\frac{d\\sigma}{dr}-\\frac{1}{2}\\frac{\\partial\\log Z_\\sigma}{\\partial\\sigma}\\left(\\frac{d\\sigma}{dr}\\right)^2 = Z_\\sigma\\frac{\\partial V_{\\rm eff}}{\\partial \\sigma}$, with boundary conditions $\\sigma'(0)=0$ and $\\sigma(\\infty)=\\sigma_F$. This is the usual Coleman bubble equation with two alterations: the effective force is multiplied by $Z_\\sigma$, and the friction receives a contribution proportional to the logarithmic derivative of $Z_\\sigma$, which can either increase or decrease friction depending on its sign. It is solved by bisection on weighted averages of the true-vacuum and barrier-crossing points, with a rescaling $V_{\\rm eff}=\\alpha^2 U$, $U(\\sigma_b)=1$, and $R=\\alpha r$ to keep the numbers moderate. For multiple fields the same equation is solved along a path of unit speed, while the perpendicular component of the equation defines a perpendicular force $N = \\left(\\frac{dx}{dr}\\right)^2\\frac{d^2\\sigma}{dx^2} - \\frac{1}{2}\\frac{d^2x}{dr^2}\\nabla_\\perp \\log Z_\\sigma - Z_\\sigma\\nabla_\\perp V_{\\rm eff}$ that drives path deformation until it vanishes. The package interpolates purely numerical potentials and renormalization factors, so no analytic expressions are required.","core_discovery":"The central discovery is that the next-to-leading-order tunneling problem can be solved numerically with two small modifications of the standard machinery. In single-field shooting, the particle-in-a-potential picture is retained but the driving force becomes $-Z_\\sigma \\partial V_{\\rm eff}/\\partial \\sigma$ and the friction is augmented by the term $-\\frac{1}{2}\\frac{\\partial \\log Z_\\sigma}{\\partial \\sigma}\\left(\\frac{d\\sigma}{dr}\\right)^2$. In multi-field tunneling, the perpendicular force that drives path deformation acquires an extra term involving the transverse gradient of $\\log Z_\\sigma$, so the renormalization factor modifies not only the bubble profile but the tunneling path itself. The package rescales the potential by its barrier height to control numerical error, handles potentials and renormalization factors that are only known numerically, and includes a supercooling-specific narrowing of the shooting search interval. With these ingredients, the paper argues, the bounce action at next-to-leading order can be computed in seconds to minutes, and the results show that $Z$ affects the action substantially even when its average value between the two vacua is normalized to 1.","pith_inferences":["Because the solver cannot run when $Z$ changes sign between the vacua, the electroweak case the Introduction singles out remains unsolved; a regularization that handles the $Z=0$ crossing is a natural next step the paper leaves open.","The large action shifts shown in the examples imply that predicted gravitational-wave spectra and nucleation temperatures could move substantially once realistic next-to-leading-order $Z$-terms are included; this is a concrete consequence worth testing in specific beyond-the-Standard-Model scenarios.","A direct test of the central claim would be to run the package on a single-field model with a non-constant $Z$ for which Eq. (2.6) has an independent solution (for example, a potential where the equation is integrable or can be solved by a completely different minimizer) and to compare the resulting action and profile; the paper validates only the $Z=1$ limit against other packages.","The path-deformation claim could be checked by comparing the package's deformed path with the path obtained by directly minimizing the next-to-leading-order action functional over paths with an independent optimizer; agreement would confirm the $Z$-dependent perpendicular force is correctly implemented."],"forward_implications":["Including $Z$ can change the bounce action by more than an order of magnitude; in the two-field example of Sec. IV.B the action shifts from 171.2 to 71283.7, so the next-to-leading-order correction is not automatically small.","The tunneling path in field space is modified by the renormalization factor, not just the field profile along the path; the $Z\\neq 1$ path differs visibly from the $Z=1$ straight-line path in the same example.","At leading order ($Z=1$), the package reproduces the actions of CosmoTransitions, FindBounce, and AnyBubble within about one percent, so the code passes the leading-order validation benchmark.","For supercooling transitions, where the barrier sits extremely close to the false vacuum, the narrowed search interval allows the solver to work at temperatures where FindBounce and other packages fail, including a test model whose vacuum separation is $10^8$ times the barrier distance.","For dark chiral phase transitions with no leading-order kinetic term, the package computes the next-to-leading-order action; full path deformation changes the action by only 0.6% relative to the lowest-potential one-dimensional path."],"supporting_citations":[{"why":"Original shooting-method bounce formalism for false-vacuum decay that the single-field solver modifies.","marker":"[44]"},{"why":"Path-deformation algorithm for multi-field bounces that this package extends to include the renormalization factor.","marker":"[66]"},{"why":"Order-by-order derivative-expansion framework that defines the next-to-leading-order effective action with kinetic coefficient $Z$.","marker":"[62]"},{"why":"Supplies the chiral phase transition effective potential and renormalization factor used as a concrete test with vanishing leading-order kinetic term.","marker":"[90]"},{"why":"FindBounce package used as the accuracy baseline for multi-field action comparisons at $Z=1$.","marker":"[92]"},{"why":"Documents the gauge dependence of vacuum decay rates that motivates the next-to-leading-order treatment with $Z$.","marker":"[83]"}],"fun_headline_variants":["Renormalization factor reshapes vacuum tunneling action","New package solves bounce with renormalization factor","Z changes tunneling paths and action in next-to-leading order","Fast numerical tunneling with renormalization included","Renormalized bounce equations solved in seconds"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The calculation assumes the wavefunction renormalization factor $Z$ is strictly positive between the false and true vacua; when $Z$ crosses zero, as it does in electroweak phase transitions, the coefficient in Eq. (2.6) blows up and the solver cannot run without a regularization the paper does not provide.","fun_headline_variants_meta":{"raw":{"variants":["Renormalization factor reshapes vacuum tunneling action","New package solves bounce with renormalization factor","Z changes tunneling paths and action in next-to-leading order","Fast numerical tunneling with renormalization included","Renormalized bounce equations solved in seconds"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000523,"raw_usage":{"total_tokens":2571,"prompt_tokens":1027,"completion_tokens":1544,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":643,"completion_tokens_details":{"reasoning_tokens":1473}},"tokens_in":643,"tokens_out":1544,"duration_ms":11237,"temperature":1.0,"reasoning_tokens":1473,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T14:28:51.818759+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Directly minimize the next-to-leading-order action functional $S=\\int dr\\,r^{D-1}\\left(\\frac{Z^{-1}}{2}\\sigma'^2 + V_{\\rm eff}\\right)$ for a single-field model with a non-constant $Z$ using an independent global method, such as fine-grid relaxation or an analytic solution where one exists, and compare the action and profile with the package's output; agreement to the claimed one-percent accuracy would confirm that Eq. (2.6) is being solved, while a larger mismatch would falsify the solver's central claim.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Original shooting-method bounce formalism for false-vacuum decay that the single-field solver modifies."},{"cited_title":"Ramsey-Musolf, Tuomas V","cited_arxiv_id":null,"evidence_quote":"Path-deformation algorithm for multi-field bounces that this package extends to include the renormalization factor."},{"cited_title":"Schwartz","cited_arxiv_id":null,"evidence_quote":"Order-by-order derivative-expansion framework that defines the next-to-leading-order effective action with kinetic coefficient $Z$."},{"cited_title":"Helmboldt, Jisuke Kubo, and Susan van der Woude","cited_arxiv_id":null,"evidence_quote":"Supplies the chiral phase transition effective potential and renormalization factor used as a concrete test with vanishing leading-order kinetic term."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"FindBounce package used as the accuracy baseline for multi-field action comparisons at $Z=1$."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Documents the gauge dependence of vacuum decay rates that motivates the next-to-leading-order treatment with $Z$."}],"review_version":1}