Pith. sign in

REVIEW 4 major objections 6 minor 59 references

The paper proposes that rapid swings in polarization angle cause geometric depolarization in both pulsars and fast radio bursts, and presents evidence from pulsars plus applications to FRB progenitors.

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

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

T0 review · deepseek-v4-flash

2026-08-01 04:27 UTC pith:MK2DFBOX

load-bearing objection A clean analytic derivation of geometric depolarization, paired with a pulsar test that is suggestive but not yet free of a noise-floor confounder — worth refereeing, with a null test as the key ask. the 4 major comments →

arxiv 2607.27622 v1 pith:MK2DFBOX submitted 2026-07-30 astro-ph.HE

Depolarization Induced by Rapid Polarization Angle Swings: A Common Feature of Pulsars and Fast Radio Bursts?

Yu-Chen Huang , Jie-Shuang Wang , Song-Bo Zhang , Zi-Gao Dai This is my paper
classification astro-ph.HE
keywords fast radio burstspulsarspolarization angledepolarizationrotating magnetospherelinear polarizationLorentz factorneutron stars
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

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

The paper argues that when the polarization angle (PA) of radio emission changes rapidly across the tiny region that contributes to the observed signal, the observed linear polarization is geometrically diluted, producing an anti-correlation between linear polarization fraction and dPA/dt. This should apply to any rotating neutron star emitter, including FRBs, regardless of the specific radiation mechanism. The authors derive an explicit relation and test it on a sample of 190 radio pulsars, finding 15 plausible anti-correlation candidates, and then apply the relation to constrain the spin periods and geometry of FRB progenitors. If correct, the relation turns polarization-angle swings into a probe of neutron-star rotation and magnetospheric structure.

Core claim

For a source whose PA is set by geometry — e.g., the projected magnetic field in a rotating magnetosphere — radiation collected within the natural beaming cone (half-angle ~1/γ) is a sum over emitters with slightly different PAs. When the PA changes by order one radian across that cone, the sum partially cancels and the linear polarization fraction drops. The paper derives an approximate closed form, Π_L ≈ |2J1(f)/f| with f = (dPA/dφ)·2/(εγ sinζ), where ζ is the angle between line of sight and spin axis and ε is a projection factor. This yields a universal anti-correlation: steeper PA swings give lower linear polarization, with essentially complete depolarization when f ≳ 4. The paper then f

What carries the argument

The central object is the observable emission cone Σ on the neutron-star unit sphere, with angular radius ~1/γ. Treating the PA ψ(x,y) as a slowly varying field over this cone, a first-order Taylor expansion turns the Stokes Q and U integrals into an oscillatory integral whose magnitude is |2J1(f)/f|, where f measures the PA swing across the cone. This Bessel-function identity is the load-bearing formula connecting the local PA gradient to the linear polarization fraction.

Load-bearing premise

The inference rests on the assumption that the observed PA reliably traces the local magnetic-field orientation in the emission region, and that the anti-correlation in pulsars is not an artifact of larger PA errors or intra-bin averaging when the signal is weakly polarized.

What would settle it

Measure, at high time resolution, a pulsar whose PA swing is steep and whose linear polarization is high; the model predicts the linear polarization fraction at the steepest swing should be approximately |2J1(f)/f| with f evaluated from the observed dPA/dt. If a source with dPA/dφ ≳ several (i.e., f≳ few) still maintains nearly full linear polarization at the same phase, the geometric depolarization is falsified. Also, a null test: if the anti-correlation disappears when restricting to bins where the PA uncertainty is small, the effect in the full sample is likely spurious.

Watch this falsifier — get emailed when new claim-graph text bears on it.

If this is right

  • For any rotating neutron-star source with geometric PA, one expects Π_L to dip where dPA/dt is largest, regardless of the detailed emission mechanism.
  • Observing the predicted dip gives an estimate of the product γ sinζ (via Eq. 25), so PA-profile morphology can constrain particle Lorentz factors.
  • If FRBs are rotating neutron stars, their PA swing rates translate into upper limits on spin period: steep swings imply sub-second periods for some one-off bursts.
  • For repeaters thought to have aligned spin and magnetic axes, the lack of depolarization can constrain the geometry and disfavor extreme alignment for sources like FRB 20240114A.
  • The relation offers a way to distinguish intrinsic geometric PA swings from propagation effects: only geometric swings should be accompanied by the characteristic Π_L dip.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The same geometric argument could be extended to radio emission from other magnetized compact objects, e.g., magnetar giant flares or γ-ray burst afterglows, predicting depolarization wherever the PA gradient across the beaming cone is steep.
  • A cleaner test would compare the predicted Π_L(f) curve point-by-point with the observed PA profiles, rather than only the rank correlation; existing data might already allow this.
  • The model's prediction is frequency-independent if the geometry is frequency-independent, so simultaneous multi-frequency polarimetry of a steep-swing pulsar could separate geometric from propagation-induced depolarization.
  • One should be cautious: the anti-correlation seen in pulsars may partially reflect that low linear polarization is measured with larger PA uncertainty, which can artificially steepen dPA/dt; a null test comparing S/N-binned subsamples would clarify.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 6 minor

Summary. The paper proposes a geometrical depolarization mechanism for pulsars and FRBs: when the polarization angle (PA) swings rapidly across the region from which radiation is collected, incoherent superposition of differently polarized contributions reduces the observed linear polarization. Starting from a general Stokes integral over a beaming cone of half-angle ~1/γ, the authors derive a closed-form relation Π_L ≈ |2J1(f)/f|, with f = (dPA/dφ)·2/(εγ sinζ) (Eqs. 21–22). They test the predicted anti-correlation between Π_L and dPA/dφ in 181 pulsars from the Wang et al. (2023) sample, reporting 15 candidates with Spearman ρs ≤ −0.7, p ≤ 0.05, of which 8 show visually clean anti-correlations. These are then used to infer Lorentz factors, and the relation is applied tentatively to several FRBs to constrain spin periods and geometry. The paper is cautious about the FRB side, but the pulsar test is the main empirical support for the central claim.

Significance. If the mechanism is correct, it provides a simple, radiation-mechanism-independent connection between PA swing rate and linear polarization fraction, with potential diagnostic power for the magnetospheric origin of FRBs. The analytical derivation in Section 2 is clean, internally consistent, and yields a falsifiable, parameter-free functional form for the depolarization once f is defined. The paper is appropriately tentative about the FRB applications. However, the empirical confirmation currently rests on a small subset of pulsars and is vulnerable to PA measurement noise and statistical multiple-comparison effects. The FRB constraints are illustrative rather than established. The theoretical part is a worthwhile contribution; the empirical part needs substantial strengthening before the broad claim can be accepted.

major comments (4)
  1. [§3.1, Eq. (23)] The pulsar test lacks a null test against PA measurement noise and finite bin-width averaging. With the adopted S/N>5 cut, σ_PA ≈ 0.1 rad, so the central-difference derivative in Eq. (23) has a noise floor σ_deriv ≈ σ_PA/(√2 Δφ) ≈ 7 rad/rad for Δφ = 0.01 rad, and is larger for narrower phase bins. The selection dPA/dφ ≥ 0.3 max(dPA/dφ) will preferentially retain noisier low-L bins, which are also often low-Π_L; this alone can produce Spearman ρs ≤ −0.7 with no intrinsic geometric depolarization. Excluding OPM jumps and showing PA fluctuations (Fig. 9) is not a null test. Since §3.2 and the FRB applications rest on this empirical anti-correlation, this is load-bearing.
  2. [§3.1, Fig. 4] The candidate selection uses thresholds (ρs ≤ −0.7, p ≤ 0.05) chosen after inspecting the data, with no multiple-comparison correction. Among 181 pulsars, about 9 p ≤ 0.05 results are expected by chance under a global null; the authors themselves note that |ρs| ≈ 0.7 appears on both sides of the distribution. A permutation or FDR analysis is needed before 15 candidates, or 8 visually clean sources, can be claimed as evidence. This matters because §3.2 uses all 15 (12 with finite ζ) to infer Lorentz factors.
  3. [§3.2, Eq. (25)] Setting εf ≃ 1 in Eq. (25) turns the Lorentz-factor 'inference' into the assumption that f = 1 at the peak, i.e., imposing the onset condition Δψ ~ 1. The resulting γ values are reparametrizations of (dPA/dφ)_peak with an order-unity (or larger) systematic uncertainty from the unknown true f. A proper treatment would fit γ, or marginalize over f, using the full Π_L–dPA/dφ profile. Without that, the γ ~ 100 prior used for the FRB predictions in §4 is not independently supported.
  4. [§4, Eqs. (26)–(30)] The FRB constraints assume the PA swing is due to rigid rotation with period P and adopt γ ~ 100 with no error budget. They also do not account for measurement uncertainties in dPA/dt or for propagation and temporal-averaging depolarization. The derived inequalities, e.g., P ≲ 0.2 s for FRB 20221022A, should be labeled illustrative; as written they may overstate the constraints. I recommend moving these to a clearly marked speculative subsection and adding explicit caveats.
minor comments (6)
  1. [§2] The symbol φ is used both for the azimuthal angle around the LOS (Eqs. 4–20) and for the spin phase in Eq. (22). Using a different symbol, e.g., ϖ, for the spin phase would remove ambiguity.
  2. [Eq. (1)] Eq. (1) is an order-of-magnitude heuristic; the text should state explicitly that it assumes ε ≈ 1 and sinζ ≈ 1, and that the equality is not a rigorous derivation.
  3. [Fig. 3] The phrase 'the measurable PA swing is able to reach a maximum at f ≈ 4' is confusing; Π_L first vanishes at f ≈ 3.83, so beyond that value the PA is effectively unmeasurable. The wording should be clarified.
  4. [§3.1] Spearman p-values are computed on phase bins that are not independent; the effective number of independent bins is smaller because of profile autocorrelation. Please state the minimum number of bins used after the N_Bin > 2 cut and consider a permutation test that preserves phase order.
  5. [Eq. (28)] In Section 4.1, 'dPA/dφ ∼ 500 deg/ms' has inconsistent units; the time derivative should be written dPA/dt. Please check Eq. (28) and the accompanying text.
  6. [Table 1] The 'Yes/No' visibility column is subjective. Provide a quantitative definition of visual anti-correlation, e.g., phase alignment of a local Π_L minimum with the local dPA/dφ maximum within a specified number of phase bins, so that the classification is reproducible.

Circularity Check

0 steps flagged

No significant circularity: the central Pi_L-dPA/dphi relation is derived from first principles and tested against external pulsar data; FRB applications are explicitly conditional.

full rationale

The central load-bearing relation, Eq. (21), Pi_L = |2J1(f)/f| with f defined in Eq. (22), is not fitted to the data it is used to explain. It follows from the standard incoherent summation of Stokes parameters (Eqs. 2-6), the assumption chi << 1 (Eq. 14), a first-order expansion of the PA field psi over the beaming cone (Eq. 16), and the resulting Bessel-function integral (Eq. 18). The geometric projection connecting f to dPA/dphi in Eq. (20) is stated as an approximation but is not itself the target prediction. The pulsar test in Sec. 3 uses an external catalogue (Wang et al. 2023), not the authors' own data, and computes Pi_L and dPA/dphi independently from L/I and a central difference of PA (Eq. 23). No parameter of Eq. (21) is adjusted to produce the reported Spearman coefficients; the selection threshold in Eq. (24) is motivated by the theory's f1/f2 ~ 0.3 ratio. The Lorentz-factor estimates in Sec. 3.2 are explicitly conditional: 'If the depolarization in the 15 pulsars shown in Fig. 5 is indeed driven by rapid PA swings, the corresponding Lorentz factors can be estimated,' and they adopt the order-of-magnitude choice 'epsilon f ~ 1'. These are applications of the model, not confirmations of it. The FRB constraints in Sec. 4 similarly state, 'The Lorentz factor here has been assumed to be roughly comparable to that in a pulsar,' and the paper later acknowledges: 'One of the major uncertainties in the application of this theory to FRBs comes from the Lorentz factor ... We have to assume gamma ~ 100.' Thus the derivation is self-contained and the applications are clearly labeled as assumption-based. A possible statistical confound (PA measurement noise producing spurious dPA/dphi at low S/N, with no null test) is a correctness risk, not circularity: it does not reduce Eq. (21) to its inputs. The only self-citations (Huang et al. 2024; Huang & Dai 2024; Liu et al. 2025b) are contextual references for rotating-magnetosphere FRB models and are not load-bearing for the derivation or the pulsar test.

Axiom & Free-Parameter Ledger

4 free parameters · 6 axioms · 0 invented entities

The derivation rests on standard Stokes superposition, beaming-cone geometry, and several simplifying assumptions about uniformity and smoothness; the pulsar test additionally assumes the RVM geometry and that PA noise does not mimic the signal. No new physical entities are introduced.

free parameters (4)
  • γ (Lorentz factor) = ~10–1779 for pulsars (Table 1); assumed ~100 for FRBs
    The depolarization scale depends on γ through f; in §3.2 it is inferred from the anti-correlation by setting ε f ≈ 1, and in §4 it is assumed γ ~ 100 to derive FRB spin-period constraints.
  • ε (projection factor sin φ_k) = ≈1 (assumed)
    ε projects the PA gradient onto the LOS sweep direction; in Eq (25) and Eq (26) the paper adopts ε f ≈ 1, effectively collapsing the product into one hand-set number.
  • f (dimensionless PA-swing parameter) = ≈1 (threshold)
    The depolarization threshold f ≈ 1 is used in Eqs (26)–(30) to translate observed dPA/dt into period limits; it is not fitted to FRB data.
  • ζ (LOS–spin-axis angle) = from RVM fits for pulsars; bounded for FRBs
    For the pulsar Lorentz-factor estimates ζ comes from Wang et al. (2023); for FRBs it is a free geometric parameter constrained only by inequalities such as Eq (30).
axioms (6)
  • domain assumption The observed radiation is the incoherent sum of many emitters, so Stokes parameters add linearly.
    Used in Eq (3); requires that emitters have random relative phases so no cross-correlation terms appear.
  • domain assumption The PA ψ(θ,φ) is a single-valued geometric function of the emission direction, tracing a preferred (e.g., magnetic-field) direction.
    Central premise stated in §2; if PA swings are instead caused by propagation effects, the geometric depolarization formula does not apply.
  • domain assumption All emitters share the same Lorentz factor γ and the observable region is a spherical cap of half-opening angle 1/γ.
    Assumed before Eq (7) to enable the closed-form integral; real emission regions likely have a spread in γ, which would modify Eq (21).
  • domain assumption The ellipticity angle χ is small (χ ≪ 1) for the bursts considered, so circular polarization is negligible.
    Used to obtain Eq (14); the analysis is deliberately restricted to bursts with intrinsically low circular polarization.
  • domain assumption The PA varies linearly across the observable region (first-order Taylor expansion of ψ(x,y)).
    Used in Eq (16); higher-order terms (multipole fields, twisting) would introduce corrections and change the exact functional form, though the anti-correlation might survive.
  • domain assumption The RVM fits of Wang et al. (2023) correctly give the geometric angles α, β, ζ for the pulsar sample.
    Used in §3.2 to convert the measured dPA/dφ into Lorentz factors via Eq (25); errors in ζ propagate strongly and three sources diverge.

pith-pipeline@v1.3.0-daily-deepseek · 15070 in / 18509 out tokens · 174718 ms · 2026-08-01T04:27:38.040869+00:00 · methodology

0 comments
read the original abstract

The polarization angle (PA) of pulsars and fast radio bursts (FRBs) provides a useful diagnostic of the magnetic fields in their emission regions and is therefore crucial for understanding their radiation and origins. Within a general geometric framework for polarized emission from a rotating neutron star, we suggest a possible anti-correlation between the degree of linear polarization $\Pi_\text{L}$ and $d\text{PA}/dt$, the time derivative of the PA, as a common feature of pulsars and FRBs. The depolarization arises from the incoherent superposition of radiation with different polarization directions within the observable part of the emission region, and is detectable only when the PA swing is steep enough. We test this conjecture using a sample of radio pulsars and find possible evidence for the expected anti-correlation in a subset of pulsars. Whether this relation holds in FRBs remains uncertain due to limited observational data. Identification of this feature would not only provide insights into the rotating magnetospheric origin of FRBs, but also place constraints on the spin periods and geometric parameters of the neutron stars that power these mysterious bursts.

Figures

Figures reproduced from arXiv: 2607.27622 by Jie-Shuang Wang, Song-Bo Zhang, Yu-Chen Huang, Zi-Gao Dai.

Figure 1
Figure 1. Figure 1: Schematic illustration showing that geometric depolarization could occur when the PA varies rapidly. Panel A: The blue curves represent the trajectories of relativistic electrons. Radiation from an electron can be observed only when its velocity direction (green arrow) lies within a cone (gray cone) with a half-opening angle of ∼ 1/𝛾 centered on the LOS (red dashed arrows). Thus, the observed radiation con… view at source ↗
Figure 2
Figure 2. Figure 2: A three-dimensional Cartesian coordinate system defined on a unit sphere centered on the neutron star (left panel) and its two-dimensional projection onto the 𝑥-𝑦 plane (right panel). The LOS unit vector 𝒏LOS is along the 𝑧-axis, and the spin axis unit vector 𝛀 is in the 𝑥-𝑧 plane. The angle between 𝛀 and 𝒏LOS is 𝜁 . The gray area indicates the observable cone Σ on the unit sphere. It has an angular size o… view at source ↗
Figure 3
Figure 3. Figure 3: The degree of linear polarization ΠL, as a function of 𝑓 , based on equation (21). Here, 𝑓 is a parameter characterizing the PA swing. The gray region indicates the range of 𝑓 where depolarization occurs and the PA remains measurable, corresponding to the transition from ΠL = 0.9 to ΠL = 0.1. The red region denotes the regime where the PA becomes essentially unmeasurable for 𝑓 ≳ 4. We then compute the degr… view at source ↗
Figure 4
Figure 4. Figure 4: Scatter plot of Spearman rank correlation coefficients 𝜌s versus p-values for 181 pulsars with sufficient data (𝑁Bin > 2). Pulsars with 𝜌s ≤ −0.7 and 𝑝 ≤ 0.05 are marked in red (middle panel). The 40 pulsars with the largest 𝑑PA/𝑑 𝜙 values are highlighted in blue (right panel). positive correlation, 𝜌s = −1 indicates a perfect anti-correlation, and 𝜌s = 0 indicates no correlation between the two variables.… view at source ↗
Figure 5
Figure 5. Figure 5: Polarization profiles of 15 pulsars identified as strongly anti-correlated candidates with 𝜌s ≤ −0.7 and 𝑝 ≤ 0.05. In each subfigure, the lower panel shows the polarization profiles: the black, red, and blue curves represent the normalized total intensity, linear polarization intensity, and circular polarization intensity, respectively. The upper panel shows the PA profiles. The 𝑆/𝑁 thresholds adopted for … view at source ↗
Figure 6
Figure 6. Figure 6: Distribution of the correlation coefficient 𝜌s for the 15 pulsars with 𝜌s ≤ −0.7 and 𝑝 ≤ 0.05. The 8 pulsars with visually apparent anti￾correlation have −1 < 𝜌s ≲ −0.8, while the remaining 7 pulsars have −0.8 ≲ 𝜌s ≤ −0.7. Among these 181 pulsars, some exhibit relatively large 𝑑PA/𝑑𝜙 values but do not show a strong anti-correlation. One possible ex￾planation is that these sources have relatively large Lore… view at source ↗
Figure 7
Figure 7. Figure 7: Distributions of max(𝑑PA/𝑑 𝜙) for pulsars selected by different 𝜌s thresholds compared to the remaining population. KS test results are provided below each panel [PITH_FULL_IMAGE:figures/full_fig_p009_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: Left Panel: Inferred Lorentz factors 𝛾 for the 12 pulsars listed in [PITH_FULL_IMAGE:figures/full_fig_p009_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: Polarization profiles of three representative pulsars, illustrating the impact of PA fluctuations on the correlation between ΠL and 𝑑PA/𝑑 𝜙. Although a visually apparent local anti-correlation can be seen, the measured correlation coefficients are not statistically significant, likely due to fluctuations in the PA data. 5 DISCUSSION We have proposed that rapid PA swings may induce depolarization in FRBs, p… view at source ↗

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

59 extracted references · 1 canonical work pages

  1. [1]

    Radiative processes in astrophysics , year =

    Rybicki, George B and Lightman, Alan P , publisher =. Radiative processes in astrophysics , year =

  2. [2]

    2023 , month = oct, number =

    Research in Astronomy and Astrophysics , title =. 2023 , month = oct, number =. arXiv , doi =:2307.10340 , keywords =

    Pith/arXiv arXiv 2023

  3. [3]

    2021 , month = apr, number =

    Science China Physics, Mechanics, and Astronomy , title =. 2021 , month = apr, number =. arXiv , doi =:2101.04907 , keywords =

    Pith/arXiv arXiv 2021

  4. [4]

    2023 , month = jul, number =

    Reviews of Modern Physics , title =. 2023 , month = jul, number =. arXiv , doi =:2212.03972 , keywords =

    Pith/arXiv arXiv 2023

  5. [5]

    1969 , month = jan, pages =

    , title =. 1969 , month = jan, pages =

    1969

  6. [6]

    2020 , month = oct, number =

    , title =. 2020 , month = oct, number =. arXiv , doi =:2011.00171 , keywords =

    Pith/arXiv arXiv 2020

  7. [7]

    2024 , month = jun, number =

    , title =. 2024 , month = jun, number =. arXiv , doi =:2401.17378 , keywords =

    Pith/arXiv arXiv 2024

  8. [8]

    2025 , month = aug, number =

    , title =. 2025 , month = aug, number =. arXiv , doi =:2504.00391 , keywords =

    Pith/arXiv arXiv 2025

  9. [9]

    2022 , month = dec, number =

    Research in Astronomy and Astrophysics , title =. 2022 , month = dec, number =. arXiv , doi =:2210.03609 , keywords =

    Pith/arXiv arXiv 2022

  10. [10]

    2020 , month = mar, number =

    , title =. 2020 , month = mar, number =. arXiv , doi =:2002.12539 , keywords =

    Pith/arXiv arXiv 2020

  11. [11]

    2024 , month = jul, number =

    , title =. 2024 , month = jul, number =. arXiv , doi =:2406.13704 , keywords =

    Pith/arXiv arXiv 2024

  12. [12]

    2024 , month = apr, number =

    , title =. 2024 , month = apr, number =. arXiv , doi =:2309.01847 , keywords =

    Pith/arXiv arXiv 2024

  13. [13]

    2019 , month = feb, number =

    , title =. 2019 , month = feb, number =. arXiv , doi =:1810.09459 , keywords =

    Pith/arXiv arXiv 2019

  14. [14]

    2025 , month = jun, number =

    , title =. 2025 , month = jun, number =. arXiv , doi =:2410.10172 , keywords =

    Pith/arXiv arXiv 2025

  15. [15]

    2018 , month = jan, number =

    , title =. 2018 , month = jan, number =. arXiv , doi =:1801.03965 , keywords =

    Pith/arXiv arXiv 2018

  16. [16]

    2023 , month = oct, number =

    , title =. 2023 , month = oct, number =. arXiv , doi =:2304.14665 , keywords =

    Pith/arXiv arXiv 2023

  17. [17]

    and Kirichenko, Aida and Pal, Arpan and Cook, Amanda M

    Mckinven, Ryan and Bhardwaj, Mohit and Eftekhari, Tarraneh and Kilpatrick, Charles D. and Kirichenko, Aida and Pal, Arpan and Cook, Amanda M. and Gaensler, B. M. and Giri, Utkarsh and Kaspi, Victoria M. and Michilli, Daniele and Nimmo, Kenzie and Pearlman, Aaron B. and Pleunis, Ziggy and Sand, Ketan R. and Stairs, Ingrid and Andersen, Bridget C. and Andre...

  18. [18]

    2022 , month = aug, pages =

    , title =. 2022 , month = aug, pages =

    2022

  19. [19]

    1975 , month = feb, pages =

    , title =. 1975 , month = feb, pages =. doi:10.1086/153393 , keywords =

    doi:10.1086/153393 1975

  20. [20]

    2025 , month = apr, number =

    , title =. 2025 , month = apr, number =. arXiv , doi =:2411.14784 , keywords =

    Pith/arXiv arXiv 2025

  21. [21]

    2025 , month = mar, number =

    , title =. 2025 , month = mar, number =. arXiv , doi =:2410.19043 , keywords =

    Pith/arXiv arXiv 2025

  22. [22]

    2025 , month = jul, number =

    , title =. 2025 , month = jul, number =. arXiv , doi =:2502.16626 , keywords =

    Pith/arXiv arXiv 2025

  23. [23]

    2025 , month = nov, number =

    , title =. 2025 , month = nov, number =. arXiv , doi =:2508.12119 , keywords =

    arXiv 2025

  24. [24]

    2025 , month = jul, pages =

    arXiv e-prints , title =. 2025 , month = jul, pages =. arXiv , doi =:2507.14708 , keywords =

    arXiv 2025

  25. [25]

    2025 , month = dec, pages =

    arXiv e-prints , title =. 2025 , month = dec, pages =. arXiv , doi =:2512.24936 , keywords =

    Pith/arXiv arXiv 2025

  26. [26]

    2001 , month = may, number =

    , title =. 2001 , month = may, number =. arXiv , doi =:astro-ph/0009266 , keywords =

    Pith/arXiv arXiv 2001

  27. [27]

    2023 , month = apr, number =

    , title =. 2023 , month = apr, number =. arXiv , doi =:2212.03988 , keywords =

    Pith/arXiv arXiv 2023

  28. [28]

    2025 , month = sep, number =

    , title =. 2025 , month = sep, number =. arXiv , doi =:2504.18176 , keywords =

    Pith/arXiv arXiv 2025

  29. [29]

    1985 , month = feb, number =

    , title =. 1985 , month = feb, number =. doi:10.1007/BF00651284 , keywords =

    work page doi:10.1007/bf00651284 1985

  30. [30]

    2016 , month = sep, number =

    , title =. 2016 , month = sep, number =. arXiv , doi =:1603.08207 , keywords =

    Pith/arXiv arXiv 2016

  31. [31]

    2016 , month = may, number =

    , title =. 2016 , month = may, number =. arXiv , doi =:1603.02014 , keywords =

    Pith/arXiv arXiv 2016

  32. [32]

    2024 , month = jan, number =

    , title =. 2024 , month = jan, number =. arXiv , doi =:2311.18487 , keywords =

    Pith/arXiv arXiv 2024

  33. [33]

    2022 , month = dec, number =

    Research in Astronomy and Astrophysics , title =. 2022 , month = dec, number =. arXiv , doi =:2210.03610 , keywords =

    Pith/arXiv arXiv 2022

  34. [34]

    2024 , month = dec, number =

    , title =. 2024 , month = dec, number =. arXiv , doi =:2310.08971 , keywords =

    Pith/arXiv arXiv 2024

  35. [35]

    2025 , month = mar, pages =

    arXiv e-prints , title =. 2025 , month = mar, pages =. arXiv , doi =:2503.12013 , keywords =

    Pith/arXiv arXiv 2025

  36. [36]

    2025 , month = may, pages =

    arXiv e-prints , title =. 2025 , month = may, pages =. arXiv , doi =:2505.14219 , keywords =

    Pith/arXiv arXiv 2025

  37. [37]

    2018 , month = oct, number =

    , title =. 2018 , month = oct, number =. arXiv , doi =:1809.03043 , keywords =

    Pith/arXiv arXiv 2018

  38. [38]

    2020 , month = nov, number =

    , title =. 2020 , month = nov, number =. arXiv , doi =:2005.10324 , keywords =

    Pith/arXiv arXiv 2020

  39. [39]

    2020 , month = nov, number =

    , title =. 2020 , month = nov, number =. arXiv , doi =:2005.10828 , keywords =

    Pith/arXiv arXiv 2020

  40. [40]

    2020 , month = nov, number =

    , title =. 2020 , month = nov, number =. arXiv , doi =:2005.11479 , keywords =

    Pith/arXiv arXiv 2020

  41. [41]

    2023 , month = jul, number =

    Science Advances , title =. 2023 , month = jul, number =. arXiv , doi =:2307.16124 , keywords =

    Pith/arXiv arXiv 2023

  42. [42]

    2024 , month = nov, number =

    , title =. 2024 , month = nov, number =. arXiv , doi =:2410.04065 , keywords =

    Pith/arXiv arXiv 2024

  43. [43]

    2026 , month = jan, pages =

    arXiv e-prints , title =. 2026 , month = jan, pages =. arXiv , doi =:2601.18865 , keywords =

    arXiv 2026

  44. [44]

    2024 , month = may, number =

    , title =. 2024 , month = may, number =. arXiv , doi =:2404.15870 , keywords =

    Pith/arXiv arXiv 2024

  45. [45]

    2026 , month = feb, pages =

    arXiv e-prints , title =. 2026 , month = feb, pages =. arXiv , doi =:2602.16409 , keywords =

    arXiv 2026

  46. [46]

    2025 , month = oct, pages =

    , title =. 2025 , month = oct, pages =. arXiv , doi =:2504.06645 , keywords =

    arXiv 2025

  47. [47]

    2026 , month = jan, number =

    , title =. 2026 , month = jan, number =. arXiv , doi =:2504.07449 , keywords =

    arXiv 2026

  48. [48]

    2026 , month = mar, pages =

    arXiv e-prints , title =. 2026 , month = mar, pages =. arXiv , eid =:2603.20663 , keywords =

    Pith/arXiv arXiv 2026

  49. [49]

    2024 , month = sep, number =

    National Science Review , title =. 2024 , month = sep, number =. arXiv , doi =:2408.03313 , keywords =

    Pith/arXiv arXiv 2024

  50. [50]

    2019 , month = aug, pages =

    , title =. 2019 , month = aug, pages =. arXiv , doi =:1906.05878 , keywords =

    Pith/arXiv arXiv 2019

  51. [51]

    2019 , month = dec, number =

    , title =. 2019 , month = dec, number =. arXiv , doi =:1904.07947 , keywords =

    Pith/arXiv arXiv 2019

  52. [52]

    1985 , month = jun, pages =

    , title =. 1985 , month = jun, pages =

    1985

  53. [53]

    2025 , month = aug, number =

    , title =. 2025 , month = aug, number =. arXiv , doi =:2508.11736 , keywords =

    Pith/arXiv arXiv 2025

  54. [54]

    2017 , month = aug, number =

    , title =. 2017 , month = aug, number =. arXiv , doi =:1703.00068 , keywords =

    Pith/arXiv arXiv 2017

  55. [55]

    2023 , month = jun, number =

    , title =. 2023 , month = jun, number =. arXiv , doi =:2302.09697 , keywords =

    Pith/arXiv arXiv 2023

  56. [56]

    2026 , month = jun, pages =

    arXiv e-prints , title =. 2026 , month = jun, pages =. arXiv , doi =:2606.19450 , keywords =

    Pith/arXiv arXiv 2026

  57. [57]

    2015 , month = mar, number =

    , title =. 2015 , month = mar, number =. arXiv , doi =:1501.00066 , keywords =

    Pith/arXiv arXiv 2015

  58. [58]

    2026 , month = apr, pages =

    arXiv e-prints , title =. 2026 , month = apr, pages =. arXiv , doi =:2604.28012 , keywords =

    Pith/arXiv arXiv 2026

  59. [59]

    2026 , month = feb, number =

    , title =. 2026 , month = feb, number =. arXiv , doi =:2601.11122 , keywords =

    arXiv 2026