Pith. sign in

REVIEW 3 major objections 6 minor 1 cited by

Testing nucleation calculations for strong phase transitions

T0 review · 3 major / 6 minor · reviewed 2026-08-08 · deepseek-v4-flash

Pith's one-line read A fully nonperturbative lattice computation of the bubble nucleation rate in a strong first-order phase transition finds that the one-loop perturbative result is too high by about 20% in $\lvert\log\Gamma\rvert$, and the tree-level result…

desk verdict A short proceedings that points to a real and worrying result—the first lattice test of nucleation in a tree-level barrier model finds the lattice rate 20% below one-loop perturbation theory—but the systematics that would settle whether this is physics or an artifact live in the companion paper. read the letter →

arxiv 2502.04185 v1 pith:V4NXI4JN submitted 2025-02-06 hep-lat hep-ph

classification hep-lathep-ph
keywords bubblenucleationfirst-orderphasetransitionlatticesimulationperturbationtheoryratemulticanonicalmethodreal-timeLangevinevolutiongravitationalwaves
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper asks whether perturbative calculations of the bubble nucleation rate survive a fully nonperturbative test. The authors simulate the three-dimensional high-temperature effective theory of a single real scalar field with a tree-level barrier — the scenario where perturbation theory should be at its best — and compare the lattice nucleation rate with tree-level, local-potential-approximation, and one-loop results. The lattice rate agrees only qualitatively: the disagreement in $\lvert\log\Gamma\rvert$ is 20% at one loop and 100% at tree level, and the nonperturbative rate lies below the one-loop prediction throughout the reweightable temperature range. Since the latent heat in the same theory matches perturbation theory to better than 1%, the failure is specific to nucleation-rate calculations, which matter for gravitational-wave signals from cosmological phase transitions.

What carries the argument

The rate computation rests on a factorisation of the nucleation rate into three pieces: the probability density of being near the critical bubble (the separatrix), an analytic Gaussian flux through the separatrix, and the fraction of near-critical configurations that actually tunnel when evolved in real time. The multicanonical weight function and the order parameter $\theta_{\mathrm{op}}=\phi^2-2A\phi$ isolate the suppressed critical-bubble peak while suppressing bulk phase fluctuations; a fourth-order symplectic integrator with momentum refresh then evolves the selected configurations forwards and backwards in a Langevin bath with damping $\gamma=1/L$, so the Monte Carlo and real-time stages agree on where the separatrix lies.

What would settle it

Run a two-loop perturbative calculation at the same benchmark point: if the predicted rate still lies above the lattice value by about 20% in $\lvert\log\Gamma\rvert$, the loop expansion is not the explanation. Alternatively, repeat the real-time evolution with a different damping coefficient or integrator timestep; if the tunneling fraction shifts enough to change $\log\Gamma$ by the observed gap, the discrepancy is an artifact of the dynamical model.

Watch

Extended reading notes

Core claim

The central claim is that, for a strong first-order phase transition with a tree-level barrier in a single-scalar theory, the perturbative nucleation rate is not quantitatively reliable. Using multicanonical sampling to build critical-bubble configurations and real-time Langevin evolution to decide which of them tunnel, the authors extract the dimensionless rate $\log(\Gamma/\lambda_3^4)$ on the lattice, continuum-extrapolate it, and compare it with analytic results. At the benchmark point, the lattice value is lower than the one-loop result by about 20% in $\lvert\log\Gamma\rvert$ and lower than the tree-level result by about 100%, with statistical errors much smaller than the perturbative renormalisation-scale bands. The paper does not claim to have identified the source of the mismatch; it concludes that higher-order perturbative calculations and further lattice studies are needed, mentioning possible extra saddle points and a potential breakdown of the saddle-point approximation as open possibilities.

Load-bearing premise

The load-bearing premise is that the real-time evolution step returns the true physical nucleation rate: if the way the simulation labels trajectories or computes the flux is wrong, the lattice rate is wrong and the discrepancy with perturbation theory is an artifact rather than a real failure of perturbation theory.

Editorial extensions

If this is right

  • If the lattice result is correct, perturbative one-loop rates overestimate bubble nucleation for strong phase transitions by about 20% in $\lvert\log\Gamma\rvert$, so bubbles would form later and the transition would supercool more than one-loop estimates suggest.
  • Tree-level rate estimates are off by roughly a factor of about $e$ in the rate itself, making them inadequate for quantitative gravitational-wave phenomenology in models with a tree-level barrier.
  • The 20% one-loop gap gives a concrete target for two-loop calculations: they must move the predicted rate downward toward the lattice value to restore confidence in the perturbative expansion.
  • Because the latent heat agrees to better than 1%, thermodynamic quantities and nucleation rates can fail independently, so nonperturbative checks of the rate itself are needed even when bulk thermodynamics looks perturbative.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • If the lattice rate is the physical one, gravitational-wave spectra computed from one-loop rates would shift: a lower rate delays percolation, which changes both the peak frequency and the amplitude of the signal predicted for LISA-era detectors.
  • The size of the gap is consistent with the paper's suggested alternatives — extra saddle points beyond the critical bubble, or a breakdown of the saddle-point expansion — and would mean the standard bounce-action framework is missing a leading-order effect, not just loop corrections.
  • A sharper test would be to split the measured rate into the flux and tunneling-fraction factors and compute each nonperturbatively, since only the tunneling fraction depends on the real-time simulation and can be checked against direct Langevin nucleation studies at higher rates.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 6 minor

Summary. This proceedings paper reports a nonperturbative lattice computation of the bubble nucleation rate in a three-dimensional scalar field theory with a tree-level barrier, which is a toy model relevant for strong cosmological phase transitions. Using multicanonical simulations to sample near-critical configurations and real-time Langevin evolution to determine the tunneling fraction, the authors obtain a continuum-infinite-volume-extrapolated lattice rate that is substantially lower than the tree-level and one-loop perturbative predictions: the disagreement in |log Γ| is quoted as 100% at tree-level and 20% at one-loop, while the latent heat agrees to better than 1%. The paper interprets this as evidence that perturbative nucleation calculations, although expected to work well for tree-level-barrier scenarios, may only be in qualitative agreement with a fully nonperturbative treatment. Technical details, including the O(a^2) improvement and the detailed extrapolations, are deferred to a companion paper [19], and the data and code are made available.

Significance. If the result is correct, it is significant for the electroweak phase transition and gravitational-wave phenomenology, because it challenges the common assumption that perturbative bubble-nucleation rates are quantitatively reliable in strong, tree-level-barrier models. The calculation is a genuine nonperturbative test: the lattice simulation is independent of the perturbative curves, so the comparison is not circular. The paper also ships reproducible data (Zenodo DOI) and open source code (scalnuc), which strengthens the reliability of the reported numbers. The main caveat is that the rate measurement relies on a factorized dynamic prescription whose validation is only partially documented in this proceedings; this is the central issue assessed in the major comments.

major comments (3)
  1. [Sec. 2.1, Eqs. (4)-(12)] The rate in Eq. (4) factorizes as P_c × (1/2)⟨flux⟩ × ⟨d⟩, where the flux is computed analytically assuming Gaussian momenta and ⟨d⟩ is measured from Langevin trajectories with damping γ = 1/L. The manuscript reports only an empirical statement that a high-order symplectic algorithm was needed so that the Monte Carlo and real-time stages agree on the separatrix; it gives no convergence tests in Δt, γ, or trajectory length, and no external validation of the factorized form. Since any bias in the flux or in ⟨d⟩ would propagate directly into Γ, the 20% and 100% discrepancies quoted in Sec. 4 are not fully supported by the evidence presented in this paper. The authors should either provide such tests or explicitly state that the discrepancy is conditional on this dynamical prescription and point to the validation in the companion paper [19].
  2. [Sec. 3, Fig. 3] The final lattice value used for the comparison in Fig. 4 is not clearly a simultaneous continuum and infinite-volume limit. The left panel of Fig. 3 shows a continuum extrapolation at fixed volume L λ_3 = 42, while the right panel shows an infinite-volume extrapolation at fixed lattice spacing a λ_3 = 1.5. Consequently, the quoted log(Γ/λ_3^4) ≈ -74.09(5) and the reweighted curves may carry undisclosed discretization or finite-volume systematics. The paper should specify which extrapolated value is used for the discrepancy percentages in Sec. 4 and what systematic error is assigned, or state that the combined continuum and infinite-volume extrapolation is deferred to [19].
  3. [Eqs. (3) and (4)] The probability density P_c in Eq. (3) depends on the arbitrary order parameter θ_op = φ^2 − 2Aφ and on the window ε, and the flux formula in Eq. (4) also depends on this choice of θ_op. The text asserts that the exact choice of ε is compensated by ⟨d⟩, but no test of invariance of the final product P_c × (1/2)⟨flux⟩ × ⟨d⟩ under changes of θ_op is presented. The single comparison using the linear order parameter φ_lin in Fig. 3 is consistent within errors, but it is not described as a systematic check. A dependence of the final rate on the arbitrary projection would invalidate the method, so the authors should report such a check or refer to a concrete verification in [19].
minor comments (6)
  1. [Sec. 1, Introduction] The phrase 'relativistic and quantum theories theories' contains a duplicated word; please correct it.
  2. [Sec. 1.1, near Eq. (2)] The discussion of O(a^2) improvement mentions a parameter κ_lat and κ_MS, but Eq. (2) does not contain a κ parameter. Please rephrase this to refer to the lattice parameters that actually appear in Eq. (2), such as σ_lat, m^2_lat, and λ_lat.
  3. [Abstract] The abstract states only that the agreement is 'qualitative'; consider adding the quantitative figures from Sec. 4 (20% at one-loop, 100% at tree-level) or a pointer to Sec. 4 so that the magnitude of the discrepancy is clear to the reader.
  4. [Eq. (4)] Please state explicitly that Γ in Fig. 4 is the rate per unit volume, and define the constant A and the normalization convention for θ_op (e.g., volume-averaged) so that the flux formula can be checked dimensionally.
  5. [Sec. 2.1] Please state how the initial momenta for the real-time evolution are sampled (for example, drawn from a Gaussian distribution) or provide a reference to the companion paper for this detail.
  6. [Figs. 3 and 4] The captions say the figures are reproduced from Ref. [19]; consider adding a sentence in the text noting that the lattice data shown here are identical to those in [19] and that the full fit details are given there.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the lattice nucleation rate is an independent nonperturbative measurement, and the perturbative comparisons are external calculations.

full rationale

The central comparison is not circular. The benchmark parameters (mu3/lambda3 = 1, sigma3/lambda3^(5/2) = -0.016687, m3^2/lambda3^2 = -0.082770, g3/lambda3^(3/2) = 0) are fixed model inputs from the xSM mapping and prior thermodynamics work [17], not quantities fitted to reproduce the target nucleation rate. The lattice rate is obtained from multicanonical probabilities and real-time trajectory fractions via the factorisation in Eq. (4), while the tree-level, one-loop and LPA curves are calculated independently and carry renormalisation-scale or complex-potential uncertainty bands. The Gaussian-flux assumption in Eq. (4) is stated explicitly rather than smuggled in, and the dynamical factorisation is the Moore-Rummukainen method, not a restatement of the perturbative prediction. Self-citations to the companion paper [19] provide O(a^2) improvement relations, continuum/infinite-volume extrapolations and final figures, but these are technical and data references within the same research programme, not an unverified premise whose content equals the claimed result; the code is archived [18] and the data are on Zenodo [28]. Possible bias in the Langevin evolution with gamma = 1/L, or in the symplectic integrator, would be a systematic/correctness risk, not circularity, because the measured rate does not reduce by construction to the perturbative inputs.

Assumptions & free parameters 3 free parameters · 6 assumptions · 0 invented entities

The central claim rests on the lattice measurement being a faithful nonperturbative evaluation of the nucleation rate. The most important items are the factorization ansatz and the real-time Langevin assumption; the extrapolation fits and matching relations are standard but are imported from the companion paper rather than derived here.

free parameters (3)
  • Continuum extrapolation coefficients b, c, d = Not stated; cubic/quartic fits in lattice spacing a
    The continuum limit log(Γ/λ3^4) at fixed Lλ3=42 is obtained from f(a)=b+c a^3 or f(a)=b+c a^3+d a^4; the fitted intercept b is the published lattice rate.
  • Infinite-volume fit coefficients b, c, m_s = Not stated; f(L)=b+c exp(-m_s L)
    Used to extrapolate the rate to infinite volume at aλ3=1.5; m_s is expected to be the screening mass.
  • Langevin damping γ = 1/L (box size L)
    Damping in trajectory evolution is chosen by hand to approximate Hamiltonian evolution and avoid lattice heating; the result may depend on this choice at finite volume.
assumptions (6)
  • domain assumption The nucleation rate factorizes as Γ V ≈ P_c^norm * 1/2 ⟨flux⟩ ⟨d⟩ (Eq. 4), with separatrix probability, flux, and tunneling fraction independent.
    The paper states the assumption of separate short and long timescale physics in Section 2.1.
  • domain assumption The momentum distribution at θc is Gaussian, giving ⟨flux⟩ = sqrt(8/(π V (θc + A^2))).
    Stated in Section 2.1 directly after Eq. (4) and used for the analytic flux factor.
  • domain assumption Real-time evolution with a Langevin thermostat and γ approaching 0 reproduces the leading-order quantum dynamics of the hot scalar field.
    Section 2.1, with citation to Aarts and Smit [22]; this is the physical interpretation of the tunneling fraction.
  • domain assumption The high-temperature physics is described by the dimensionally reduced 3D effective theory Eq. (1) with long-wavelength modes.
    Section 1.1 states this assumption explicitly.
  • domain assumption The O(a^2) improved lattice action and the lattice-continuum matching relations of the companion paper [19] give the correct continuum parameters.
    Section 1.1 refers to the Supplemental Material of Ref. [19]; this matching is not derived in the present paper.
  • domain assumption The maximum of the free energy between phases corresponds to a critical bubble rather than slabs or cylinders.
    Section 2 states this requires sufficiently large cubic lattices; this is the basis for identifying the separatrix with the critical bubble.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Testing nucleation calculations for strong phase transitions." pith.science (2026). https://pith.science/paper/V4NXI4JN

@misc{pith2026250204185,
  author       = {Pith},
  title        = {Pith review of: Testing nucleation calculations for strong phase transitions},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/V4NXI4JN}},
  note         = {Machine review of arXiv:2502.04185}
}
abstract

Accurate calculations of the nucleation rate $\Gamma$ for first order phase transitions are important for determining their observable consequences in particle physics and cosmology. Perturbative calculations are often used, but they are incomplete and should be tested against fully nonperturbative lattice simulations. We simulate nucleation on the lattice in a scalar field theory with a tree-level barrier, a scenario which should be well described by perturbation theory. Our computation of the nucleation rate, however, only shows qualitative agreement with the perturbative result. This motivates further study of nucleation on the lattice and to higher orders in perturbation theory.

Figures

Figures reproduced from arXiv: 2502.04185 by the authors.

Figure 1
Figure 1. Free energy as the function of the order parameter at some temperature below 𝑇𝑐. The metastable and stable phases are separated by an exponentially suppressed mixed phase. The maximum of the free energy in this mixed phase corresponds to the critical bubble. Configurations near the critical bubble - the separatrix - are drawn from a narrow range 𝜖 and act as initial conditions for the real time dynamics of the syste… view at source ↗
Figure 2
Figure 2. Schematic illustration of trajectories starting from near-critical bubble configurations, drawn from the narrow range 𝜖. Configurations are evolved backwards and forwards in time from the starting point and then combined into a full trajectory that is used to determine whether a given configuration tunnels or not. The system is evolved forwards and backwards for a given number of timesteps, or until the field leaves… view at source ↗
Figure 3
Figure 3. At left, the continuum extrapolation for fixed volume 𝐿𝜆3 = 42, along with one point computed with the linear order parameter (denoted 𝜙lin). Given our 𝑂(𝑎 2 ) improvement, we fit to cubic 𝑓 (𝑎) = 𝑏 + 𝑐𝑎3 and quartic 𝑓 (𝑎) = 𝑏 + 𝑐𝑎3 + 𝑑𝑎4 , noting also that our largest lattice spacing is comparable to the inverse screening mass; we thus exclude it from the cubic fit. At right, an extrapolation to infinite volume for… view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: The nucleation rate as a function of temperature for the tree-level result, the local potential approximation (LPA), and one-loop perturbative results. For the tree-level and one-loop results the uncertainty bands are given by varying the renormalisation scale, while f…

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Consistent Thermal Resummation and Phase Transitions with 2PI Methods

    hep-ph 2026-08 conditional novelty 6.0 of 10

    A 2PI-Hartree effective potential for two mixing scalars is renormalized and used to show that self-consistent thermal resummation can substantially alter predicted phase transition strengths and gravitational wave spectra.

Reference graph

Works this paper leans on

28 extracted references · 9 canonical work pages · cited by 1 Pith paper

  1. [19]

    Gould, A

    O. Gould, A. Kormu and D.J. Weir,Nonperturbative test of nucleation calculations for strong phase transitions, Phys. Rev. D(2025, in press) [2404.01876]

  2. [1]

    Rel.26(2023) 5 [2204.05434]

    LISA Cosmology Working Group collaboration,Cosmology with the Laser Interferometer Space Antenna,Living Rev. Rel.26(2023) 5 [2204.05434]

  3. [2]

    Langer,Theory of the condensation point,Annals Phys.41(1967) 108

    J.S. Langer,Theory of the condensation point,Annals Phys.41(1967) 108

  4. [3]

    Langer,Statistical theory of the decay of metastable states,Annals Phys.54 (1969) 258

    J.S. Langer,Statistical theory of the decay of metastable states,Annals Phys.54 (1969) 258

  5. [4]

    Coleman,The Fate of the False Vacuum

    S.R. Coleman,The Fate of the False Vacuum. 1. Semiclassical Theory,Phys. Rev. D15 (1977) 2929

  6. [5]

    Affleck,Quantum Statistical Metastability,Phys

    I. Affleck,Quantum Statistical Metastability,Phys. Rev. Lett.46 (1981) 388

  7. [6]

    A.D.Linde, DecayoftheFalseVacuumatFiniteTemperature ,Nucl.Phys.B 216(1983)421

  8. [7]

    Zenesini, A

    A. Zenesini, A. Berti, R. Cominotti, C. Rogora, I.G. Moss, T.P. Billam et al.,False vacuum decay via bubble formation in ferromagnetic superfluids,Nature Phys.20 (2024) 558 [2305.05225]

Show all 28 references
  1. [8]

    Low Temp

    QUEST-DMCcollaboration, A-B Transition in Superfluid3He and Cosmological Phase Transitions,J. Low Temp. Phys.215 (2024) 495 [2401.07878]

  2. [9]

    Alford, H

    M.G. Alford, H. Feldman and M. Gleiser,Thermal activation of metastable decay: Testing nucleation theory,Phys. Rev. D47 (1993) R2168

  3. [10]

    Alford and M

    M.G. Alford and M. Gleiser,Metastability in two-dimensions and the effective potential, Phys. Rev. D48(1993) 2838 [hep-ph/9304245]

  4. [11]

    Borsanyi, A

    S. Borsanyi, A. Patkos, J. Polonyi and Z. Szep,Fate of the classical false vacuum, Phys. Rev. D 62(2000) 085013 [hep-th/0004059]

  5. [12]

    Batini, A

    L. Batini, A. Chatrchyan and J. Berges,Real-time dynamics of false vacuum decay,Phys. Rev. D109 (2024) 023502 [2310.04206]

  6. [13]

    Pîrvu, A

    D. Pîrvu, A. Shkerin and S. Sibiryakov,Thermal false vacuum decay in (1+1) dimensions: Evidence for nonequilibrium dynamics, Int. J. Mod. Phys. A39 (2024) 2445007 [2408.06411]

  7. [14]

    Moore and K

    G.D. Moore and K. Rummukainen,Electroweak bubble nucleation, nonperturbatively,Phys. Rev. D63(2001) 045002 [hep-ph/0009132]

  8. [15]

    Moore, K

    G.D. Moore, K. Rummukainen and A. Tranberg,Nonperturbative computation of the bubble nucleation rate in the cubic anisotropy model,JHEP 04 (2001) 017 [hep-lat/0103036]

  9. [16]

    Gould, S

    O. Gould, S. Güyer and K. Rummukainen,First-order electroweak phase transitions: A nonperturbative update, Phys. Rev. D106 (2022) 114507 [2205.07238]

  10. [17]

    Gould,Real scalar phase transitions: a nonperturbative analysis, JHEP 04(2021) 057 [2101.05528]

    O. Gould,Real scalar phase transitions: a nonperturbative analysis, JHEP 04(2021) 057 [2101.05528]. 9 Testing nucleation calculations for strong phase transitions David J. Weir

  11. [18]

    Scalnuc release 2.1.0

    O. Gould, A. Kormu and D.J. Weir, “Scalnuc release 2.1.0.” SWHID swh:1:rel:94596986a4cf3bfd61ed75e34de3fe46a66d4753; origin=https://bitbucket.org/og113/scalnuc.git; visit=swh:1:snp:85d8b3bf7b4550676db3ca4913273cd89ea0a40f, 2025

  12. [20]

    B.A.BergandT.Neuhaus, Multicanonicalensemble: ANewapproachtosimulatefirstorder phase transitions, Phys. Rev. Lett.68(1992) 9 [hep-lat/9202004]

  13. [21]

    Rummukainen, R

    K. Rummukainen, R. Seppä and D.J. Weir,Resolving the critical bubble inSU(8) deconfinement transition,PoS LATTICE2024(2025) 434 [2501.17593]

  14. [22]

    Aarts and J

    G. Aarts and J. Smit,Classical approximation for time dependent quantum field theory: Diagrammatic analysis for hot scalar fields, Nucl. Phys. B511 (1998) 451 [hep-ph/9707342]

  15. [23]

    Andreassen, D

    A. Andreassen, D. Farhi, W. Frost and M.D. Schwartz,Precision decay rate calculations in quantum field theory,Phys. Rev. D95 (2017) 085011 [1604.06090]

  16. [24]

    Ekstedt,Bubble nucleation to all orders, JHEP 08(2022) 115 [2201.07331]

    A. Ekstedt,Bubble nucleation to all orders, JHEP 08(2022) 115 [2201.07331]

  17. [25]

    Pîrvu, M.C

    D. Pîrvu, M.C. Johnson and S. Sibiryakov,Bubble velocities and oscillon precursors in first-order phase transitions,JHEP 11 (2024) 064 [2312.13364]

  18. [26]

    Pîrvu, A

    D. Pîrvu, A. Shkerin and S. Sibiryakov,Thermal False Vacuum Decay Is Not What It Seems, 2407.06263

  19. [27]

    Gould and T.V.I

    O. Gould and T.V.I. Tenkanen,On the perturbative expansion at high temperature and implications for cosmological phase transitions, JHEP 06(2021) 069 [2104.04399]

  20. [28]

    Finaldataforpaper A nonperturbative test of nucleation calculations for strong phase transitions

    O.Gould,A.KormuandD.J.Weir,“Finaldataforpaper A nonperturbative test of nucleation calculations for strong phase transitions.” DOI 10.5281/zenodo.10891523, 2024. 10

Pith tools

Reviewed August 8, 2026 · model on record in the stance chip above.