Pith. sign in

REVIEW 2 major objections 6 minor 35 references

Representing Equations of State With Strong First-Order Phase Transitions

T0 review · 2 major / 6 minor · reviewed 2026-08-07 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read Spectral representations of neutron-star equations of state remain slightly more accurate than piecewise analytic ones at equal parameter count, even when the equation of state has a strong first-order phase transition.

desk verdict Useful, clearly written numerical test with a legitimate but non-fatal caveat: single-start Levenberg-Marquardt fits leave the spectral-vs-piecewise ranking slightly underdetermined. read the letter →

arxiv 2506.06201 v1 pith:OQZBNS4E submitted 2025-06-06 nucl-th astro-ph.HEgr-qc

Lee Lindblom , Steve M. Lewis , Fridolin Weber This is my paper
classification nucl-thastro-ph.HEgr-qc
keywords equationofstateneutronstarsfirst-orderphasetransitionspectralrepresentationpiecewiseanalyticcausalityChebyshevpolynomialsquarkmatter
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

This paper asks whether parametric fits of the neutron-star equation of state stay accurate when the true equation of state contains a strong first-order phase transition of the kind that may separate hadronic from quark matter. It constructs causal spectral and piecewise-analytic representations of the ACB4 and ACB5 model equations of state without telling the fit where the transition is, and it measures the error with the dimensionless residual $\chi$ defined on the energy-density-versus-pressure curve. The central finding is that spectral representations are generally somewhat more accurate than piecewise-analytic representations with the same number of parameters, and that both families converge at about the same power-law rate. This directly answers a standing criticism that spectral methods could fail for transitions strong enough to destabilize the neutron-star sequence while creating a disconnected branch of stable quark-core stars.

What carries the argument

The load-bearing device is the velocity function $\Upsilon(p) = c^2\,d\epsilon/dp - 1 = (c^2-v^2)/v^2$, which is non-negative exactly when the equation of state is both thermodynamically stable and causal. Any parametric fit is built by choosing a non-negative $\Upsilon(p,\upsilon_a)$ and integrating $\epsilon(p) = \epsilon_{\min} + (p-p_{\min})/c^2 + c^{-2}\int_{p_{\min}}^p \Upsilon(p')\,dp'$, so every representation automatically satisfies causality and stability. The paper compares two families: a spectral family in which $\Upsilon$ is an exponential of a weighted Chebyshev-polynomial sum in $\log p$, and a piecewise family in which $\Upsilon$ is a power law on logarithmically spaced pressure intervals. A two-zone variant adds three parameters that place and size the phase-transition discontinuity and fits separate representations below and above it. The error measure $\chi$ is minimized with Levenberg-Marquardt, and the comparison of minimal $\chi$ versus $N_{\rm parms}$ carries the argument.

What would settle it

Run the same parameter optimization with a global search or many random restarts on a suite of discontinuous model equations of state; if for any such case a piecewise-analytic fit achieves a substantially smaller minimal $\chi$ than the spectral fit at the same $N_{\rm parms}$, the claim that spectral representations are generally somewhat more accurate would be falsified.

Watch

Extended reading notes

Core claim

On the paper's own terms, the discovery is that a causal Chebyshev spectral representation of the high-density equation of state remains a perfectly workable tool for equations of state with strong hadron-quark first-order transitions: for the discontinuous $\Delta P = 0$ ACB4 and ACB5 models, the fit error $\chi$ decreases monotonically from about 0.5 at one parameter to about 0.06 at ten parameters, and the spectral fits are generally somewhat more accurate than piecewise-analytic fits with the same parameter count. Both families converge roughly as $1/N_{\rm parms}$, which is the best rate expected for discontinuous functions when the discontinuity location and size are unknown. The paper also shows that one-zone representations beat two-zone representations that split the pressure range at the phase transition for every case tested, up to eleven parameters. The inference drawn is that the earlier criticism of spectral representations for missing instability-triggering phase transitions is not supported by these tests.

Load-bearing premise

The fits for every representation and parameter count are assumed to reach the global minimum of the error measure $\chi$; if some Levenberg-Marquardt runs stop in local minima, the accuracy ranking between spectral and piecewise families could change.

Editorial extensions

If this is right

  • Spectral representations can be used with confidence in inverse stellar-structure studies even when the true equation of state contains a strong first-order phase transition.
  • For equations of state with unknown transition locations and strengths, roughly $1/N_{\rm parms}$ convergence is the expected ceiling for both representation families.
  • With eleven or fewer parameters, one-zone representations are more efficient than two-zone representations, so two-zone fits are not needed for near-future observational analyses.
  • Smoother mixed-phase transitions with larger $\Delta P$ are represented more accurately than the discontinuous $\Delta P=0$ case by both methods.
  • Adding parameters to a piecewise fit does not always lower the error monotonically, because the phase-transition point's proximity to a subdomain boundary varies with $N_{\rm parms}$.

Reading between the lines

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

  • Beyond the paper: because the one-zone fits remain accurate across the discontinuity, inverse stellar-structure solvers can likely use the same spectral representation for hybrid hadron-quark stars without needing to detect the transition location in advance.
  • Beyond the paper: the roughly $1/N_{\rm parms}$ ceiling suggests that adding parameters to one-zone fits is inefficient; a more promising accuracy gain would come from a representation that builds in the expected discontinuity structure while keeping the parameter count low.
  • Beyond the paper: the comparison tested only fits with eleven or fewer parameters, so the authors' expectation that two-zone representations eventually win could be checked directly by extending the curves to larger $N_{\rm parms}$.
  • Beyond the paper: nothing in the spectral construction uses specific properties of the ACB4 and ACB5 models, so the accuracy ranking is likely to transfer to other equations of state with strong first-order transitions.
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

2 major / 6 minor

Summary. The paper studies how accurately equations of state with strong first-order phase transitions can be represented by causal Chebyshev-spectral and piecewise-analytic parametric forms, using the ACB4 and ACB5 hadron-quark model EOS with and without a mixed-phase region. It introduces one-zone representations over the full pressure range and two-zone representations that split at the phase transition, and compares their accuracy as measured by the logarithmic error χ with up to eleven parameters. The central claims are that spectral representations are generally somewhat more accurate than piecewise analytic representations at equal parameter number, that both converge roughly as 1/Nparms, and that one-zone representations are more accurate than two-zone representations for the parameter counts tested.

Significance. If the claims are correct, the paper provides a useful answer to a criticism that spectral EOS representations fail for equations of state with instability-triggering first-order phase transitions. The test problems are physically motivated and the error measure is clearly defined. The paper also raises a practical question about whether two-zone representations that explicitly locate the phase transition are worth their extra parameters. The main strengths are the clean formulation of the comparison, the use of independent benchmark EOS from the literature, and the explicit attention to causality and stability through the velocity-function parameterization. However, the conclusions rest on the assumption that the reported Levenberg-Marquardt fits reach global minima, and the two-zone comparison is not performed on equal terms because the three phase-transition parameters are fixed to exact values. Both issues need to be addressed before the ranking claims can be taken as established.

major comments (2)
  1. [Section IV] The Levenberg-Marquardt minimization described in Section IV is a local optimization method, and the manuscript does not report its initial guesses or any multi-start or global search. Since the central claim in Section VI and Fig. 11 is that spectral representations are generally more accurate than piecewise analytic representations with the same number of parameters, and since Figs. 9-10 show non-monotonic χ(Nparms) for piecewise fits, the observed differences could reflect local minima rather than the expressive power of the representation families. Please add multi-start or sensitivity analysis, or otherwise justify that the reported χ values are close to the global minima.
  2. [Section III C / Section V] The two-zone comparison in Fig. 14 and the related conclusion in Section VI count ppt, εpt, and Υpt as parameters, but those values are fixed to the exact ACB4 phase-transition values rather than fitted. A comparison at equal total Nparms is therefore not a comparison of representations with the same number of adjustable parameters, and the statement that one-zone representations are "more efficient" than two-zone representations for Nparms ≤ 11 is not supported as stated. Please either fit these parameters as part of the minimization or explicitly frame the result as a comparison with a two-zone ansatz that is given the exact phase-transition properties.
minor comments (6)
  1. [Section V, Fig. 15] The claim that both representations converge roughly as 1/Nparms is supported only by a single reference curve χ = 0.5/Nparms drawn through the data. A quantitative fit of the power-law exponent, with error bars, would make the convergence-rate comparison more convincing.
  2. [Section V] The text says the upper pressure limit pmax corresponds to the central pressure of the maximum-mass quark star for ACB4(ΔP), but the same pmax is used for the ACB5 fits. Please clarify whether this range is also appropriate for ACB5 or state the actual range used for ACB5.
  3. [Section V, Eq. (13)-(14)] The figure captions and text refer to "χ defined in Eq. (14)", but Eq. (14) defines the derivative of χ²; the error measure is defined by Eq. (13). Please correct the cross-references.
  4. [Figure captions 9-10] The captions for Figs. 9 and 10 describe χ as a function of "the number of spectral parameters" for piecewise analytic representations; the wording should be "the number of parameters".
  5. [Section V] There is a typographical error in the sentence "have smaller errors than than the representation" that should read "smaller errors than the representation".
  6. [General] No data tables or code are provided for the fitted parameter values or the underlying EOS tables. Making the data available would improve reproducibility, especially for readers who want to test alternative initial guesses or global optimization methods.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the central accuracy comparison is an external benchmark against independent ACB4/ACB5 equation-of-state tables, and no claimed prediction reduces to a fitted input or to a self-citation chain.

full rationale

The paper's central claim is that spectral representations are generally somewhat more accurate than piecewise analytic representations with the same number of parameters for equations of state with strong first-order phase transitions, and that both converge at roughly the same power-law rate. This is established by fitting each representation family to the independent ACB4 and ACB5 equation-of-state tables from Refs. [27] and [29] and comparing the minimized residuals chi defined in Eq. (13). The target equations of state are not constructed from either representation family being tested, so the benchmark is external to the claimed result. The only quantities taken from the exact equations of state are the two-zone phase-transition parameters (ppt, eps_pt, and Upsilon_pt), which are openly declared in Section V and actually make the two-zone representations easier by giving them exact phase-transition information; despite this, the one-zone representations win, so this input cannot be a source of circular advantage. The prior self-citations, such as Refs. [19], [22], and [23], supply the representation formulas and earlier convergence evidence, but the present numerical tests are performed anew against independent tables and do not rely on those citations as the proof of the main result. The potential concern that Levenberg-Marquardt fits might find local rather than global minima is a numerical correctness and robustness issue, not a circularity: even a suboptimal fit would be a fair, if possibly pessimistic, measurement for that representation family, and it does not make the comparison equivalent to its inputs by construction. No step in the derivation defines the claimed accuracy in terms of the representation parameters, and no fitted parameter is relabeled as a prediction. Therefore, no circular step is present.

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

The paper's central claims rest on a small set of numerical and modeling assumptions. The most consequential are the choice of the ACB4 and ACB5 model equations of state as ground truth, the χ error metric, and the unstated assumption that nonlinear least-squares fits are globally optimal. No new physical entities are introduced.

free parameters (3)
  • Two-zone phase transition parameters (ppt, εpt, Υpt) = ppt=1.01221731e35 erg/cm^3, εpt=1.01995577e15 g/cm^3, Υpt=0.586578885
    Set to the exact values from the ACB4 equation of state rather than fitted or inferred. They enter the two-zone representations and bias the comparison in favor of two-zone, yet one-zone still wins.
  • Υ_min = 13.7751914
    Boundary value of the velocity function at p_min, evaluated from the low-density equation of state to avoid an artificial phase-transition discontinuity. Not fitted, but chosen, and it affects all representations.
  • Representation coefficients υ_a (spectral) and υ_k (piecewise) = Various, not tabulated
    The adjustable parameters optimized by Levenberg-Marquardt for each equation of state and parameter count. The paper reports only the resulting χ, not the coefficient values, so the fits cannot be reproduced without re-running the optimization.
assumptions (6)
  • domain assumption ACB4 and ACB5 hadron-quark equations of state, and their ΔP generalizations, are valid representatives of strong first-order phase transitions in neutron-star matter.
    Used as ground truth for all accuracy tests; conclusions about convergence are drawn from these models only.
  • standard math The velocity-function representation Υ = c^2/v^2 - 1 with Υ ≥ 0 ensures thermodynamic stability and causality.
    Invoked in Section III to construct causal representations; standard relativistic relation v^2 = dp/dε.
  • domain assumption The logarithmic error measure χ defined in Eq. (13) adequately captures the accuracy needed for astrophysical inference.
    Used to rank representations; no connection to observables like mass, radius, or tidal deformability is demonstrated.
  • ad hoc to paper Levenberg-Marquardt minimization converges to the global minimum of χ for each representation.
    Assumed in Section IV; no multi-start or global optimization is described.
  • standard math Chebyshev spectral expansions of discontinuous functions converge algebraically, with power-law rather than exponential rates.
    Cited from Boyd [25] and used to interpret the observed 1/N convergence.
  • standard math The Oppenheimer-Volkoff equations correctly describe the stellar structure sequences used to select the pressure range and identify instability.
    Used in Section II to compute mass-radius curves and to choose p_min and p_max.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Representing Equations of State With Strong First-Order Phase Transitions." pith.science (2026). https://pith.science/paper/OQZBNS4E

@misc{pith2026250606201,
  author       = {Pith},
  title        = {Pith review of: Representing Equations of State With Strong First-Order Phase Transitions},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/OQZBNS4E}},
  note         = {Machine review of arXiv:2506.06201}
}
read the original abstract

Parametric representations of the high-density nuclear equation of state are used in constructing models for interpreting the astrophysical observations of neutron stars. This study explores how accurately equations of state with strong first-order phase transitions can be represented using spectral or piecewise analytic methods that assume no {\it{a priori}} knowledge of the location or the strength of the phase transition. The model equations of state used in this study have phase transitions strong enough to induce a gravitational instability that terminates the sequence of stable neutron stars. These equations of state also admit a second sequence of stable stars with core matter that has undergone this strong first-order phase transition (possibly driven by quark deconfinement). These results indicate that spectral representations generally achieve somewhat higher accuracy than piecewise analytic representations having the same number of parameters. Both types of representation show power-law convergence at approximately the same rate.

Figures

Figures reproduced from arXiv: 2506.06201 by the authors.

Figure 1
Figure 1. FIG. 1: This figure illustrates the ACB4(∆ [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 4
Figure 4. FIG. 4: This figure illustrates in more detail the regions of [PITH_FULL_IMAGE:figures/full_fig_p003_4.png] view at source ↗
Figure 5
Figure 5. FIG. 5: This figure illustrates the mass-radius curves obtai [PITH_FULL_IMAGE:figures/full_fig_p004_5.png] view at source ↗
Figures from the paper (8 more)
Figure 6
Figure 6. Figure 6: FIG. 6: This figure illustrates the mass-radius curves obtai [PITH_FULL_IMAGE:figures/full_fig_p004_6.png]
Figure 7
Figure 7. Figure 7: FIG. 7: This figure illustrates the equation of state error me [PITH_FULL_IMAGE:figures/full_fig_p006_7.png]
Figure 8
Figure 8. Figure 8: FIG. 8: This figure illustrates the equation of state error me [PITH_FULL_IMAGE:figures/full_fig_p006_8.png]
Figure 11
Figure 11. Figure 11: FIG. 11: This figure illustrates the error measures [PITH_FULL_IMAGE:figures/full_fig_p007_11.png]
Figure 10
Figure 10. Figure 10: FIG. 10: This figure illustrates the equation of state error [PITH_FULL_IMAGE:figures/full_fig_p007_10.png]
Figure 14
Figure 14. Figure 14: FIG. 14: This figure illustrates the equation of state accura [PITH_FULL_IMAGE:figures/full_fig_p008_14.png]
Figure 13
Figure 13. Figure 13: FIG. 13: This figure illustrates the fractional errors, [PITH_FULL_IMAGE:figures/full_fig_p008_13.png]
Figure 15
Figure 15. Figure 15: FIG. 15: This figure illustrates the convergence rates of the [PITH_FULL_IMAGE:figures/full_fig_p009_15.png]

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

35 extracted references · 26 canonical work pages

  1. [1]

    Lindblom and N

    L. Lindblom and N. M. Indik, Phys. Rev. D 86, 084003 (2012)

    work page 2012

  2. [2]

    Lindblom and N

    L. Lindblom and N. M. Indik, Phys. Rev. D 89, 064003 (2014)

    work page 2014

  3. [3]

    Lindblom, AIP Conference Proceedings 1577, 153 (2014)

    L. Lindblom, AIP Conference Proceedings 1577, 153 (2014)

    work page 2014

  4. [4]

    Lindblom, Phys

    L. Lindblom, Phys. Rev. D 98, 043012 (2018)

    work page 2018

  5. [5]

    ¨Ozel and P

    F. ¨Ozel and P. Freire, Ann. Rev. Astron. and Astroph. 54, 401 (2016)

    work page 2016

  6. [6]

    T. E. Riley et al., Astrophys. J. Lett. 887, L21 (2019)

    2019

  7. [7]

    M. C. Miller et al., Astrophys. J. Lett. 887, L24 (2019)

    2019

  8. [8]

    M. C. Miller et al., Astrophys. J. Lett. 918, L28 (2021)

    2021

Show all 35 references
  1. [9]

    Biswas, Astrophys

    B. Biswas, Astrophys. J. 921, 63 (2021)

    2021

  2. [10]

    Vinciguerra et al., Astrophys

    S. Vinciguerra et al., Astrophys. J. 961, 62 (2024)

    2024

  3. [11]

    Salmi et al

    T. Salmi et al. (2024), arXiv:2406.14466

    2024 arXiv

  4. [12]

    Choudhury et al., Astrophys

    T. Choudhury et al., Astrophys. J. Lett. (2024), arXiv:2407.06789

    2024 arXiv

  5. [13]

    (2024), arXiv:2407.11153

    K.Chatziioannou et al. (2024), arXiv:2407.11153

    2024

  6. [14]

    S. De, D. Finstad, J. M. Lattimer, D. A. Brown, E. Berger, and C. M. Biwer, Phys. Rev. Lett. 121, 091102 (2018)

    2018

  7. [15]

    S. De, D. Finstad, J. M. Lattimer, D. A. Brown, E. Berger, and C. M. Biwer, Phys. Rev. Lett. 121, 259902 (2018)

    2018

  8. [16]

    Chatziioannou, Gen

    K. Chatziioannou, Gen. Relativ. Gravit. 52, 109 (2020)

    2020

  9. [17]

    J. S. Read, B. D. Lackey, B. J. Owen, and J. L. Friedman, Phys. Rev. D79, 124032 (2009)

    2009

  10. [18]

    Lindblom, Phys

    L. Lindblom, Phys. Rev. D 82, 103011 (2010)

    2010

  11. [19]

    Lindblom, Phys

    L. Lindblom, Phys. Rev. D 97, 123019 (2018)

    2018

  12. [20]

    Annala, T

    E. Annala, T. Gorda, A. Kurkela, J. Nattila, and A. Vourinen, Nature Physics 16, 907 (2020)

    2020

  13. [21]

    Lindblom, Phys

    L. Lindblom, Phys. Rev. D 105, 063031 (2022)

    2022

  14. [22]

    Lindblom and T

    L. Lindblom and T. Zhou, Phys. Rev. D 110, 083030 (2024)

    2024

  15. [23]

    Lindblom, Phys

    L. Lindblom, Phys. Rev. D 110, 043018 (2024)

    2024

  16. [24]

    Lindblom and T

    L. Lindblom and T. Zhou, Phys. Rev. D 111, 063024 (2025)

    2025

  17. [25]

    J. P. Boyd, Chebyshev and Fourier Spectral Methods (Dover Publications, 1999), 2nd ed

    1999

  18. [26]

    Lindblom, Phys

    L. Lindblom, Phys. Rev. D 58, 024008 (1998)

    1998

  19. [27]

    Paschalidis, K

    V. Paschalidis, K. Yagi, D. Alvarez-Castillo, D. B. Blaschke, and A. Sedrakian, Phys. Rev. D 97, 084038 (2018)

    2018

  20. [28]

    P. L. Reed Essick and D. E. Holz, Phys. Rev. D 101, 063007 (2020)

    2020

  21. [29]

    Blaschke, D

    D. Blaschke, D. E. Alvarez-Castillo, A. Ayriyan, H. Grigorian, N. K. Largani, and F. Weber, in Topics on Strong Gravity: A Modern View on Theories and Experi- ments, edited by C. A. Z. Vasconcellos (World Scientific, 2020), chap. 7, pp. 207–256

    2020

  22. [30]

    N. K. Glendenning, Phys. Rev. D 46, 1274 (1992)

    1992

  23. [31]

    N. K. Glendenning, Physics Reports 342, 392 (2001)

    2001

  24. [32]

    Abgaryan, D

    V. Abgaryan, D. Alvarez-Castillo, A. Ayriyan, D. Blaschke, and H. Grigorian, Universe 4, 94 (2018)

    2018

  25. [33]

    J. R. Oppenheimer and G. M. Volkoff, Phys. Rev. 55, 374 (1939)

    1939

  26. [34]

    L. D. Landau and E. M. Lifshitz, Fluid Mechanics (Perg- amon Press, 1959)

    1959

  27. [35]

    W. H. Press, S. A. Teukolsky, W. T. Vetterling, and B. P. Flannery, Numerical Recipes in FORTRAN (Cambridge University Press, Cambridge, England, 1992), 2nd ed

    1992

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.