Pith. sign in

REVIEW 3 major objections 5 minor 1 cited by

Uncovering Correlations and Biases in Parameter Inference from Neutron-Star Pulse Profile Modeling

T0 review · 3 major / 5 minor · reviewed 2026-08-11 · deepseek-v4-flash

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

Pith's one-line read Systematic errors in pulse-profile radius measurements reach about 1.5 km for neutron stars whose polar-cap temperature falls below the observed X-ray band.

desk verdict A useful reparameterization and a plausible warning about cool polar caps, but the headline bias is a self-consistency check of the analytic model rather than a calibrated systematic. read the letter →

arxiv 2412.12283 v1 pith:HPF5X6PK submitted 2024-12-16 astro-ph.HE gr-qc

Tong Zhao , Dimitrios Psaltis , Feryal Ozel , Elif Beklen This is my paper · ORCID
classification astro-ph.HEgr-qc
keywords neutronstarspulseprofilemodelingcompactnessX-rayprofilesNICERparameterdegeneracybeamingMarkovchainMonteCarlo
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 tries to establish that two previously under-appreciated effects corrupt pulse-profile inferences of neutron-star mass and radius: hidden degeneracies among geometric and lensing parameters, and the mismatch between the observed X-ray energy band and the temperature of the polar caps. Using an analytic light-bending model, it shows that with isotropic surface emission the data constrain only three combinations of four geometric parameters, leaving one combination free; non-isotropic beaming breaks this degeneracy. For stars with effective temperatures near 0.15 keV, whose spectral peak falls at the edge or outside the NICER band, the inferred compactness can be systematically off by roughly 0.05, translating to radius errors up to about 1.5 km. This matters because the NICER sources with published radii include such cool stars, so the reported radii may carry systematic uncertainty comparable to their formal errors.

What carries the argument

The load-bearing tool is an analytic Schwarzschild-plus-Doppler pulse profile model built on the approximate light-bending relation $\cos\alpha \approx u + (1-u)\cos\psi$, which lets the flux be written in closed form and expanded in Fourier harmonics. From that expansion the paper constructs a reparameterization — $q = u + (1-u)\cos\theta\cos\zeta$, $s = u - (1-u)\cos\theta\cos\zeta$, $p = (1-u)\sin\theta\sin\zeta$, $T_\infty = T'\sqrt{1-u}$, and $A = dS/D^2$ — that removes the strongest correlations and makes MCMC sampling efficient. The Fourier expressions show directly that for isotropic beaming the flux depends on the geometric parameters only through $p/q$, $p/s$, and $qA$, explaining the degeneracy analytically. The beaming factor $h(E',T')$ is fit to atmosphere models as a quadratic in $E'/kT'$, and the observed energy range then enters through the temperature dependence of the blackbody spectrum.

What would settle it

Generate synthetic NICER-like observations with a fully numerical ray-tracing code, including time delays and finite spot sizes, for stars with $T = 0.15$ keV and a range of geometries, then fit them with the paper's analytic model and MCMC pipeline; if the distribution of $u_{\rm fit} - u_{\rm syn}$ remains centered with width near 0.05, the energy-band bias is confirmed, whereas if the width drops to the formal statistical error, the claimed bias is an artifact of the analytic approximation.

Watch

Extended reading notes

Core claim

The central claim is that the systematic uncertainty in the inferred compactness $u = 2GM/Rc^2$ for neutron stars with effective temperatures $\lesssim 0.15$ keV can be as large as about 0.05, and up to about 0.1 depending on geometry, even when the total photon number is held fixed and the fitting model is the same one used to generate the data. This corresponds to a radius error of roughly 1.5 km for a typical neutron star. The paper also claims that isotropic beaming ($h=0$) creates a complete degeneracy in which only $p/q$, $p/s$, and $qA$ can be constrained, while $q$ is unconstrained, and that non-isotropic beaming is what allows $q$ to be measured. The implication is that a priori knowledge of the surface beaming and an energy range covering the spectral peak are prerequisites for trustworthy radius measurements from pulse profiles.

Load-bearing premise

The bias numbers come from fitting synthetic data that were generated with the same analytic model used for the fits; the model is checked against numerical ray tracing to within a few percent in flux, but a few percent model error is comparable to the roughly 0.05 compactness bias, so the magnitude or sign of the real-world bias could shift if the true light bending, time delays, or beaming differ.

Editorial extensions

If this is right

  • For NICER-like observations of stars with $T \lesssim 0.15$ keV, the compactness inferred from a single pulse profile can be off by up to about 0.05, so any single-star radius claim should carry a systematic error budget of order 1 km, not just the formal MCMC error.
  • Because the bias appears without any dependence on geometry or visibility class, averaging over multiple stars or geometries will not remove it; only observing a wider energy band or a hotter source would.
  • With isotropic beaming the parameter $q$ is unconstrained, meaning that pulse profiles alone cannot fix the combination of compactness, inclination, and spot colatitude; a theory of surface beaming is needed to break the degeneracy.
  • The new parameterization using $q$, $p/q$, $s/p$, $T_\infty$, and $Aq$ should make posterior sampling far more efficient in future analyses, potentially avoiding the multimodal sampling failures reported in earlier NICER reanalyses.

Reading between the lines

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

  • If the claimed bias is real, then the published NICER radii for PSR J0030+0451 and PSR J0740+6620, both cool sources, may need corrections of order a kilometer, with the direction depending on the geometry of the spots.
  • A natural extension the paper leaves implicit is to generate the synthetic data with a fully numerical ray-tracing code and fit them with the analytic model; if the ~0.05 compactness bias persists, it is a property of the energy band, whereas if it disappears, the analytic light-bending approximation is the source.
  • The same reparameterization logic could be applied to more complex spot geometries, where non-antipodal or unequal spots introduce additional degeneracies; a Fourier analysis would likely reveal new invariant combinations of parameters.
  • The paper's focus on the spectral peak suggests a concrete mission-level test: for a given target, verify that the observed energy band brackets the peak of the time-averaged spectrum before trusting a radius measurement from pulse profiles.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 5 minor

Summary. The paper constructs an analytic Schwarzschild+Doppler model for X-ray pulse profiles from a neutron star with two antipodal hot spots, introduces a reparametrization (q, s, p, T∞, A) intended to reduce parameter correlations, and studies, via MCMC fits to synthetic NICER-like data, how beaming and the observed energy range affect parameter recovery. The main quantitative result is that for cool effective temperatures (T ≈ 0.15 keV), where the spectral peak falls at or outside the low-energy edge of the NICER band, the recovered compactness scatters by up to about 0.05–0.1 depending on geometry and noise realization (Section 6, Figures 7–8). The analytic model is compared to numerical ray tracing in Appendix A, with flux residuals below about 0.5% for a small spot at 1 Hz and about 2.5% at 200 Hz.

Significance. The reparametrization in Section 3 is a genuinely useful contribution: it makes degeneracies analytically transparent and can accelerate posterior exploration in pulse-profile modeling. The demonstration that non-isotropic beaming breaks the q-degeneracy and that the location of the spectral peak relative to the observed band strongly affects parameter constraints is relevant to interpreting current NICER results. The beaming coefficients are anchored to external Salmi et al. (2020) atmosphere models rather than tuned to the paper's own outputs, which is a methodological strength. The main limitation is that the headline quantitative bias is measured under the same analytic forward model used to generate the data, so its real-world applicability depends on the model-validation step that is currently only performed at the flux level.

major comments (3)
  1. [§6, Figs. 7–8] The central claim that cool stars incur compactness systematics as large as ~0.05–0.1 is obtained by generating and fitting synthetic pulse profiles with the same analytic S+D model (Section 4). This measures the scatter of the posterior mode under the assumed forward model, not the systematic error that would arise from model misspecification. Appendix A validates the analytic model only at the level of flux residuals (≲2.5% at 200 Hz, ~10% at 500 Hz) for a limited set of geometries and spot sizes, and does not propagate those residuals into parameter posteriors. Because a flux error of ~2.5% is comparable to the claimed compactness shift (Δu ≈ 0.05 is ~11% of u = 0.45), the magnitude and even the sign of the effect could change if the true light bending, time delays, finite spot size, or beaming differ from the analytic treatment. I request a validation in which numerical ray-tracing profiles (or at least an approximate model-error injection at the 2–3% level) are fitted with the analytic model and the ufit−usyn distributions are recomputed; alternatively, the claims in Section 6 should be explicitly restricted to 'bias under the analytic model.'
  2. [§2.2 vs. Appendix A] The main analysis assumes δD = γ = 1 and neglects Doppler effects (Section 2.2), yet Appendix A validates the analytic model against numerical ray tracing at 200 Hz and 500 Hz using an analytic calculation in which the Doppler factor was retained. Consequently the model used for the Section 5.3 bias study is not exactly the model validated in Appendix A. Since the primary NICER targets spin at 170–270 Hz and the Doppler factor enters as the fourth power in flux, the paper should either include Doppler effects in the synthetic data and the fitting model, or explicitly state that the bias study excludes them and justify that the omission does not change the conclusions.
  3. [§5.1, Fig. 3; §6, Fig. 8] The single-realization shift Δu = −0.02 reported in Section 5.1 is a finite-noise fluctuation of the posterior mode, not a demonstrated systematic bias. In the same vein, the distributions in Figure 8 are centered near zero (median 0.003 and −0.002 for the two temperatures), so the quantity being reported is an inflation of the scatter of the point estimate rather than a persistent offset. The text should consistently distinguish 'increased scatter' from 'systematic bias' when discussing ufit−usyn, since the abstract and summary currently use 'systematics' and 'biases' interchangeably.
minor comments (5)
  1. [Table 2] In the row for s, the range is given as '-1 < q < 1'; it should presumably read '-1 < s < 1.'
  2. [§5.3] The scan is described as 'for 10° ≤ θ ≤ 80° and 10° ≤ θ ≤ 80°'; the second variable should be ζ, the spot colatitude.
  3. [Fig. 8] The solid 'cumulative distributions' are computed over a prescribed grid of θ and ζ values rather than draws from a specified prior. Calling them cumulative distributions and quoting quantiles assumes a weighting that is not described; please specify the exact grid and weighting used.
  4. [Appendix A, Figs. 9–11] The residual panels in Figures 9–11 use different vertical scales; please state the scales explicitly in the captions so that the reader can compare the magnitudes of the residuals.
  5. [§4] The paper does not report convergence diagnostics for the MCMC chains (e.g., Gelman-Rubin statistics, chain lengths, acceptance rates) or a statement on code availability; adding these would improve reproducibility.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: degeneracy claims are derived algebraically from the paper's own Fourier expansions, beaming coefficients are fitted to external Salmi et al. (2020) models, and the compactness-bias numbers are Monte Carlo outputs under a stated forward model rather than fitted inputs.

full rationale

The paper's derivation chain is non-circular. The analytic flux model in Section 2 comes from the external Beloborodov (2002) and Poutanen & Beloborodov (2006) S+D framework; the one self-cited correction (Psaltis et al. 2019, the 1/gamma factor) is explicitly set to unity for the calculations and is therefore not load-bearing. The 'weakly degenerate' parameters q, s, p, T_inf, and A introduced in Table 2 are algebraic combinations of the physical parameters, and the claimed degeneracies follow from the paper's own Fourier amplitudes (eqs. 16-35), not from any fitted value. The beaming-function coefficients (Table 3) are obtained by fitting Equation (B1) to the external Salmi et al. (2020) atmosphere models and then fixed; they are inputs to the forward model, not quantities the paper claims to predict. The headline compactness bias of ~0.05 for cool stars (Section 6, Figures 7-8) is the result of generating synthetic data with the analytic model and fitting those data with the same model; this is a controlled Monte Carlo estimate of estimator bias under a stated generative model, so the output is not equivalent by construction to an input (the fitted u is not a parameter used to generate the data). Appendix A compares the analytic model to numerical ray tracing and explicitly reports residuals of <=2.5% at 200 Hz and up to ~10% at 500 Hz; this is an acknowledged model-validity limitation, not a circular step, because the bias calculation is conditional on the analytic model and the numerical check is not a fitted input. The caveat that the real-world bias could differ if the forward model is inaccurate is a model-validity concern, not circular reasoning.

Assumptions & free parameters 2 free parameters · 6 assumptions · 0 invented entities

The central claims rest on the analytic S+D model, the antipodal two-spot geometry, the blackbody plus beaming parameterization, and the choice of synthetic data. No new physical entities are introduced. The beaming coefficients are fitted to external atmosphere models, and the synthetic-data choices are settings of the experiment.

free parameters (2)
  • Beaming fit coefficients a, b, c = a=0.0500, b=0.3695/0.3106/0.0955, c=-0.00976 for delta=1,2,3
    Fitted to Salmi et al. (2020) beaming functions at Teff=0.4 keV in Appendix B. These set the beaming model used to interpret the impact of beaming, though the qualitative conclusions use the extremes h=0 and h=1.
  • Effective temperature of beaming fits Teff = 0.4 keV
    Single temperature used for the beaming fits; scaling to other temperatures relies on the Eddington-Barbier approximation in Appendix B.
assumptions (6)
  • domain assumption Schwarzschild+Doppler spacetime with Beloborodov light-bending approximation (eq. 3)
    Used throughout; validated in Appendix A to less than 2.5% for the NICER spin range, but it remains an approximation.
  • domain assumption Two antipodal hot spots of equal size and temperature
    Section 3; the paper notes non-antipodal spots would exacerbate degeneracies, so the results are a lower bound on the biases.
  • domain assumption Blackbody spectrum with beaming factor normalization eq. (7)
    Section 2.3 and Appendix B; the beaming parameterization is fitted to Salmi et al. (2020) models and extended to other temperatures via the Eddington-Barbier scaling.
  • domain assumption Doppler effects neglected (delta_D = gamma = 1)
    Section 2.2; stated to not affect degeneracy exploration, but it changes the energy mapping and phase dependence of the flux.
  • domain assumption Spot treated as infinitesimally small
    Section 2.2 and Appendix A; finite spot size introduces less than 2.5% differences but can smooth occultation.
  • domain assumption Flat-top priors over model parameters
    Section 4; the choice of prior ranges can affect marginalized posteriors and biases.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Uncovering Correlations and Biases in Parameter Inference from Neutron-Star Pulse Profile Modeling." pith.science (2026). https://pith.science/paper/HPF5X6PK

@misc{pith2026241212283,
  author       = {Pith},
  title        = {Pith review of: Uncovering Correlations and Biases in Parameter Inference from Neutron-Star Pulse Profile Modeling},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/HPF5X6PK}},
  note         = {Machine review of arXiv:2412.12283}
}
read the original abstract

Modeling of X-ray pulse profiles from millisecond pulsars offers a promising method of inferring the mass-to-radius ratios of neutron stars. Recent observations with NICER resulted in measurements of radii for three neutron stars using this technique. In this paper, we explore correlations between model parameters and the degree to which individual parameters can be inferred from pulse profiles, using an analytic model that allows for an efficient and interpretable exploration. We introduce a new set of model parameters that reduce the most prominent correlations and allow for an efficient sampling of posteriors. We then demonstrate that the degree of beaming of radiation emerging from the neutron star surface has a large impact on the uncertainties in the inferred model parameters. Finally, we show that the uncertainties in the model parameters for neutron stars for which the polar cap temperature falls outside of the NICER energy range are significantly degraded.

Figures

Figures reproduced from arXiv: 2412.12283 by the authors.

Figure 1
Figure 1. The geometry of localized emission on the surface of a neutron star that we consider in this paper. A hot spot is located on the surface at a colatitude ζ and a phase angle ϕ. The observer is looking at the star center at an inclination angle θ. The unit surface normal vector through the center of the spot is nˆ and the unit vector along the line of sight is ˆk. The angle between nˆ and the line of sight is ψ, while… view at source ↗
Figure 2
Figure 2. The beaming factor h as a function of photon energy (in units of the effective temperature of the atmosphere), for neutron￾star atmospheres bombarded by relativistic leptons with a power￾law energy spectrum of index δ. The filled circles are the results of fitting eq. (7) to the calculations of Salmi et al. (2020), while the solid curves are quadratic fits to these data points. The dashed curve shows the approximate… view at source ↗
Figure 3
Figure 3. The posterior of fitting synthetic pulse profile data in the energy range 0.3 keV−1.5 keV, using a model described in terms of the physical parameters shown in [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗
Figures from the paper (7 more)
Figure 4
Figure 4. Figure 4: The posterior of fitting the same synthetic data as those in [PITH_FULL_IMAGE:figures/full_fig_p008_4.png]
Figure 5
Figure 5. Figure 5: The posterior of fitting synthetic data with the same neutron-star parameters as in [PITH_FULL_IMAGE:figures/full_fig_p009_5.png]
Figure 6
Figure 6. Figure 6: The time-averaged, synthetic NICER countrate spectrum for two neutron stars with different temperatures, normalized such that the total number of photons in each spectrum is the same. At T = 0.15 keV, the rapid drop of the spectrum at the higher energies increases subs…
Figure 8
Figure 8. Figure 8: Cumulative histograms of the differences ufit − usyn be￾tween the fitted compactness ufit = 2GM/Rc2 and its ground-truth value for two configurations with usyn = 0.45. The blue curves cor￾responds to a cooler neutron star (T = 0.15 keV) while the red lines curves to a …
Figure 9
Figure 9. Figure 9: Comparison of the pulse profiles from a small spot on a slowly rotating neutron star, generated using the analytic model (solid curves) and numerical ray-tracing calculations (filled circles) for two different beaming functions. The half opening angle of the spot is 5◦…
Figure 10
Figure 10. Figure 10: Comparison of the pulse profiles generated using the analytic model (solid curve) and numerical ray-tracing calculations (filled circles) when the half opening angle of the spot in the numerical model is set to 20◦ . In the left panel, the observer inclination and spo…
Figure 11
Figure 11. Figure 11: Same as in Fig.9 but for stellar spin frequencies of f = 200 Hz in the left panel and f = 500 Hz in the right panel. Ignoring the effects of time delays introduces a ≲ 2.5% error in the simulated pulse profiles for the slow spin frequencies of the primary NICER target…

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. Constraining Pulsar Radiative Geometry via Multi-wavelength Modeling

    astro-ph.HE 2026-07 conditional novelty 6.0 of 10

    Joint X-ray pulse-profile and radio RVM modeling can locate the polarization focus and test whether propagation effects dominate pulsar radio polarization.

Reference graph

Works this paper leans on

36 extracted references · 20 canonical work pages · cited by 1 Pith paper

  1. [1]

    P., Abbott, R., Abbott, T

    Abbott, B. P., Abbott, R., Abbott, T. D., et al. 2018, PhRvL, 121, 161101 Bauböck, M., Psaltis, D., & Özel, F. 2015a, ApJ, 811, 144 —. 2015b, ApJ, 811, 144 —. 2019, ApJ, 872, 162

    work page 2018

  2. [2]

    Beloborodov, A. M. 2002, ApJL, 566, L85

    work page 2002

  3. [3]

    2013, ApJ, 762, 96

    Bogdanov, S. 2013, ApJ, 762, 96

    2013

  4. [4]

    E., & Rybicki, G

    Bogdanov, S., Grindlay, J. E., & Rybicki, G. B. 2008, ApJ, 689, 407

    2008

  5. [5]

    B., & Grindlay, J

    Bogdanov, S., Rybicki, G. B., & Grindlay, J. E. 2007, ApJ, 670, 668

    2007

  6. [6]

    J., Ho, W

    Bogdanov, S., Dittmann, A. J., Ho, W. C. G., et al. 2021, ApJL, 914, L15

    work page 2021

  7. [7]

    M., Romani, R

    Braje, T. M., Romani, R. W., & Rauch, K. P. 2000, ApJ, 531, 447

    work page 2000

  8. [8]

    M., Leahy, D., & Campbell, S

    Cadeau, C., Morsink, S. M., Leahy, D., & Campbell, S. S. 2007, ApJ, 654, 458

    2007

Show all 36 references
  1. [9]

    2024, ApJL, 971, L20

    Choudhury, D., Salmi, T., Vinciguerra, S., et al. 2024, ApJL, 971, L20

    2024

  2. [10]

    T., Fonseca, E., Ransom, S

    Cromartie, H. T., Fonseca, E., Ransom, S. M., et al. 2020, Nature Astronomy, 4, 72

    2020

  3. [11]

    Hessels, J. W. T. 2010, Nature, 467, 1081 Güver, T., Özel, F., Cabrera-Lavers, A., & Wroblewski, P. 2010a, ApJ, 712, 964 Güver, T., Wroblewski, P., Camarota, L., & Özel, F. 2010b, ApJ, 719, 1807

    2010

  4. [12]

    O., Edmonds, P

    Heinke, C. O., Edmonds, P. D., Grindlay, J. E., et al. 2003, ApJ, 590, 809

    2003

  5. [13]

    O., Rybicki, G

    Heinke, C. O., Rybicki, G. B., Narayan, R., & Grindlay, J. E. 2006, ApJ, 644, 1090

    2006

  6. [14]

    E., Özel, F., & Psaltis, D

    Lockhart, W., Gralla, S. E., Özel, F., & Psaltis, D. 2019, MNRAS, 490, 1774

    2019

  7. [15]

    1978, Stellar atmospheres

    Mihalas, D. 1978, Stellar atmospheres

    1978

  8. [16]

    C., Lamb, F

    Miller, M. C., Lamb, F. K., Dittmann, A. J., et al. 2019, ApJL, 887, L24 —. 2021, ApJL, 918, L28

    2019

  9. [17]

    1983, ApJ, 270, 119

    Morrison, R., & McCammon, D. 1983, ApJ, 270, 119

    1983

  10. [18]

    M., & Leahy, D

    Morsink, S. M., & Leahy, D. A. 2011, ApJ, 726, 56 Özel, F., Baym, G., & Güver, T. 2010, PhRvD, 82, 101301 Özel, F., & Freire, P. 2016, ARA&A, 54, 401 Özel, F., Güver, T., & Psaltis, D. 2009, ApJ, 693, 1775 Özel, F., Psaltis, D., Arzoumanian, Z., Morsink, S., & Bauböck, M. 2016...

    2011

  11. [19]

    R., Ftaclas, C., & Cohen, J

    Pechenick, K. R., Ftaclas, C., & Cohen, J. M. 1983, ApJ, 274, 846

    1983

  12. [20]

    2020, A&A, 640, A24

    Poutanen, J. 2020, A&A, 640, A24

    2020

  13. [21]

    Poutanen, J., & Beloborodov, A. M. 2006, MNRAS, 373, 836

    2006

  14. [22]

    2014, ApJ, 792, 87

    Psaltis, D., & Özel, F. 2014, ApJ, 792, 87

    2014

  15. [23]

    2019, Physical Set-Up and Approximations for Light Curve Modeling, doi:10.5281/zenodo.3534254

    Psaltis, D., Ozel, F., Baubock, M., & Chan, C.-K. 2019, Physical Set-Up and Approximations for Light Curve Modeling, doi:10.5281/zenodo.3534254

    2019 doi

  16. [24]

    2014, ApJ, 787, 136

    Psaltis, D., Özel, F., & Chakrabarty, D. 2014, ApJ, 787, 136

    2014

  17. [25]

    2022, ApJ, 928, 55

    Psaltis, D., Özel, F., Medeiros, L., et al. 2022, ApJ, 928, 55

    2022

  18. [26]

    A., Özel, F., & Psaltis, D

    Raithel, C. A., Özel, F., & Psaltis, D. 2018, ApJL, 857, L23 —. 2021, ApJ, 908, 103

    2018

  19. [27]

    E., Watts, A

    Riley, T. E., Watts, A. L., Bogdanov, S., et al. 2019, ApJL, 887, L21

    2019

  20. [28]

    E., Watts, A

    Riley, T. E., Watts, A. L., Ray, P. S., et al. 2021, ApJL, 918, L27

    2021

  21. [29]

    2018, A&A, 618, A161

    Salmi, T., Nättilä, J., & Poutanen, J. 2018, A&A, 618, A161

    2018

  22. [30]

    F., Nättilä, J., & Poutanen, J

    Salmi, T., Suleimanov, V . F., Nättilä, J., & Poutanen, J. 2020, A&A, 641, A15

    2020

  23. [31]

    2023, ApJ, 956, 138

    Salmi, T., Vinciguerra, S., Choudhury, D., et al. 2023, ApJ, 956, 138

    2023

  24. [32]

    2024, arXiv e-prints, arXiv:2406.14466

    Salmi, T., Choudhury, D., Kini, Y ., et al. 2024, arXiv e-prints, arXiv:2406.14466

    2024 arXiv

  25. [33]

    W., Lattimer, J

    Steiner, A. W., Lattimer, J. M., & Brown, E. F. 2010, ApJ, 722, 33 —. 2013, ApJL, 765, L5

    2010

  26. [34]

    N., & Arons, J

    Timokhin, A. N., & Arons, J. 2013, MNRAS, 429, 20

    2013

  27. [35]

    L., et al

    Vinciguerra, S., Salmi, T., Watts, A. L., et al. 2023, ApJ, 959, 55 —. 2024, ApJ, 961, 62

    2023

  28. [36]

    2024, ApJ, submitted

    Zhao, T., Psaltis, D., & Özel, F. 2024, ApJ, submitted

    2024

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.