Pith. sign in

REVIEW 4 major objections 5 minor 36 references

Collisionless plasma turbulence still builds larger magnetic fields, but more slowly than MHD predicts, because kinetic effects break self-similar decay.

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 ↗

In collisionless 2D plasma turbulence, inverse transfer persists but decays about 40–50% slower in time than MHD predicts, with exponents that depend on magnetization.

T0 review reviewed 2026-08-02 challenge →

load-bearing objection A credible PIC study showing collisionless nonhelical inverse transfer is slower than MHD predicts, but the headline exponents rest on a short fit window with no error bars. the 4 major comments →

arxiv 2607.13406 v1 pith:7XBPPCSO submitted 2026-07-15 astro-ph.HE

Inverse Transfer in Non-helical 2D Collisionless Magnetic Turbulence: Island-Merger Picture with Kinetic Effects

Yangyang Cai , Hongzhe Zhou , Yosuke Mizuno This is my paper
classification astro-ph.HE
keywords magnetic inverse transfercollisionless plasmadecaying turbulenceparticle-in-cell simulationpressure anisotropyisland mergerself-similaritymagnetization
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 reading

The paper asks whether magnetic inverse transfer—the growth of magnetic energy toward larger scales—still works when the plasma is collisionless, as in many astrophysical environments. Using particle-in-cell simulations of decaying nonhelical 2D turbulence, the authors find that the MHD-style invariant B²ξ_b² roughly holds, so the energy–scale relation survives. However, the decay exponents are systematically smaller than the MHD island-merger prediction (p<1, q<1/2) and vary with initial magnetization. The magnetic spectrum is not self-similar: the spectral peak migrates faster than the integral scale, and Larmor-scale structure and pressure anisotropy appear. If correct, astrophysical extrapolations from MHD decay scalings overestimate how fast collisionless plasmas grow large-scale magnetic fields.

Core claim

In 2.5D particle-in-cell simulations of freely decaying nonhelical pair-plasma turbulence, magnetic inverse transfer persists: the in-plane magnetic energy B² and the magnetic integral scale ξ_b maintain B²ξ_b²≈const, and fitted power laws B²~t^{-p}, ξ_b~t^q satisfy p≈2q. But the equal-island MHD expectation p=1, q=1/2 is not met. Instead, p and q are lower in every magnetization run (p=0.58–0.77, q=0.27–0.39) and depend systematically on σ0. The spectral peak moves toward lower wavenumber faster than ξ_b grows (q_peak≈0.43–0.56), and the post-peak spectrum contains two breaks near the Larmor radius, with time-dependent slopes. The authors argue this broken self-similarity—caused by pressure

What carries the argument

The island-merger picture: small magnetic islands coalesce into larger ones, conserving 2D flux so that B²ξ_b²≈const, while each merger takes a reconnection time τ≈ξ_b/(ε_rec v_A). In the self-similar MHD limit these relations force B²~t^{-1} and ξ_b~t^{1/2}. The paper tests this machinery in a collisionless setting and shows that pressure anisotropy (double-adiabatic, with p⊥<p∥) and kinetic-scale spectral structure break the single-length-scale assumption, yielding slower decay. The conservation-like constraint B²ξ_b²≈const is retained empirically, but the local-to-global time-scale link fails.

Load-bearing premise

The conclusion that decay is slower than MHD depends on fitting a single power law over one chosen time window (20 ≤ t/l0 ≤ 100); if the true evolution is not yet a settled power law in that window, the fitted exponents and the 'p<1, q<1/2' result could shift with the interval chosen.

What would settle it

Re-run the same simulations but fit b²(t) and ξ_b(t) over substantially different time windows (or use a time-dependent exponent, e.g., a running power law). If the fitted p and q drift noticeably or approach p=1, q=1/2 in a later window, the claim of systematically slower-than-MHD decay would be undermined.

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

If this is right

  • Astrophysical estimates that use MHD decay scalings to predict how fast small magnetic seeds grow to large coherence lengths will overestimate the growth rate in collisionless environments such as the solar wind, pulsar-wind nebulae, and intergalactic plasma.
  • The energy–scale invariant B²ξ_b²≈const can still serve as a useful diagnostic for collisionless inverse transfer even when the system is not self-similar.
  • Because the decay exponents vary with initial magnetization, there is no single universal power law for collisionless inverse transfer; models must account for the guide-field strength or magnetization.
  • The spectral peak and the integral scale evolve at different rates, so different definitions of 'coherence length' must be kept distinct when comparing observations or simulations.
  • Pressure anisotropy consistently approaches the firehose condition in the runs, correlating with both the reduced effective tension and the broken self-similarity.

Where Pith is reading between the lines

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

  • If this result carries to 3D, the guide-field-to-plasma pressure ratio could become a practical control parameter for predicting large-scale field growth in gamma-ray-burst and pulsar-wind contexts—an extension the authors leave implicit.
  • The decoupling of k_peak and ξ_b suggests that different observational tracers (e.g., synchrotron polarization decorrelation versus Faraday-rotation measures) may effectively measure different magnetic length scales, which could reconcile apparently conflicting coherence-length estimates.
  • A testable extension would be to check whether the fitted exponents converge to the MHD values as σ0→∞ and as numerical resolution increases, which would indicate that the kinetic slow-down is a finite-magnetization, finite-Larmor-radius effect.
  • The broad initial island-area distribution (a ~S^-2 tail) may explain part of the slowdown; a focused experiment starting from a monodisperse island hierarchy would isolate this population effect from the pressure-anisotropy effect.
Share X Bluesky LinkedIn Reddit HN

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 / 5 minor

Summary. The paper studies inverse transfer in decaying, non-helical, collisionless (pair-plasma) turbulence using 2.5D PIC simulations at five guide-field magnetizations σ0 = 0.25, 1, 4, 16, 64. The authors report that the in-plane magnetic energy B² and integral scale ξ_b approximately satisfy B²ξ_b² ≃ const, and that power-law fits B²∝t^{-p}, ξ_b∝t^q give p≈2q but p<1 and q<1/2 in all cases, i.e. slower decay than the MHD island-merger prediction (p=1, q=1/2). They also find the spectral peak migrates faster than ξ_b (q_peak ≈ 0.43–0.56 vs q ≈ 0.27–0.39), and that the post-peak magnetic spectrum has two breaks near the Larmor scale and evolves with time, indicating broken self-similarity. The slower decay is attributed to a broad initial island-area distribution and to pressure-anisotropy-reduced magnetic tension. The authors argue that MHD-based decay-time scalings may overestimate the coherence-growth rate in collisionless astrophysical plasmas.

Significance. If the quantitative claim of systematically slower collisionless inverse transfer (p<1, q<1/2, with magnetization dependence) is robust, it would correct a widely used MHD extrapolation in magnetogenesis, GRB afterglow, and cluster-plasma contexts. The paper’s strengths include a high-resolution PIC setup (8192² grid) with a dedicated convergence test for σ0=16 (up to 16392² and 256 ppc), five magnetizations, BIC-selected piecewise spectral fitting, and an explicit dependence of exponents on σ0 that is novel for kinetic inverse transfer. The analysis also honestly acknowledges its own limitations (Section 7-8: initial-population effect not quantified, model deliberately simple, not a closed astrophysical model). The central qualitative message—that kinetic effects modify the decay-time scaling while preserving the energy–scale relation—is plausible and timely.

major comments (4)
  1. [§4, Fig. 3] The headline exponents p and q are obtained from power-law fits over a single interval 20≤t/l0≤100, spanning only about 0.7 decades in time, with no reported uncertainties and no sensitivity to the interval endpoints. Since Section 5 (Fig. 7) shows that the spectral indices s2 and s3 'still increase at the end of the simulations', the spectral shape is still evolving inside the fitting window; the fitted p and q may therefore be effective transient exponents rather than asymptotic scaling laws. Please provide bootstrap/least-squares uncertainties, vary the fit window (e.g., start at 30 or 50, end at 80 or 120), and show that the p<1, q<1/2 conclusion is stable.
  2. [§5, Fig. 7; §4] The manuscript argues that a single power law 'is an adequate description of the main decay stage' (Section 4) while simultaneously showing that the normalized spectrum is not self-similar (Section 5). These two statements are in tension: if the post-peak spectrum is still evolving, the decay may not have settled into a unique scaling regime. The paper should either identify a criterion for when the asymptotic regime is reached or explicitly frame the exponents as effective, finite-time exponents and discuss how much the MHD comparison is affected.
  3. [§6, §6.1, §6.2] The two proposed mechanisms for the slow decay are not quantitatively separated. Section 6.1 states that the initial-population effect 'does not measure how much of each fitted exponent is caused by the initial distribution', and Section 6.2 offers only a correlation between T_rms and σ0 without a causal model. Given that the abstract and conclusions attribute the σ0-dependence to kinetic effects, the paper should either perform a controlled test (e.g., compare with a run initialized with identical islands) or soften the causal language to 'consistent with' rather than 'attributed to'.
  4. [§3, Fig. 2] The conservation-like relation b²ξ_b²≃const is inferred from the slope of b² vs ξ_b in Fig. 2. The slope is quoted as α≈2 but no uncertainty is given, and the relation is then used to interpret p≈2q. While this is not circular (Fig. 2 is independent of the time fits in Fig. 3), the stability of α across the fitting range and across realizations should be reported to justify 'approximate conservation' as a quantitative constraint.
minor comments (5)
  1. [General / typography] There are numerous typographical issues, e.g., 'Alv´ en' instead of 'Alfvén', missing spaces in 'T ransfer', 'T able', 'S −2' etc. A careful proofreading pass is needed.
  2. [§2, Eq. (13)] The definition l0 = L/N2 is clear, but the relation to the excited-mode band (N1=33, N2=64) could be stated more explicitly; the reference scale is associated with the highest initialized wavenumber, which is fine but should be noted as a choice.
  3. [§4, Fig. 4] In Fig. 4 the ratio q/p is shown to be ~0.5, but the individual error bars are absent; adding error bars to p, q, and q_peak would make the figure much more informative.
  4. [§7] The astrophysical implications section is speculative but appropriately hedged. However, the claim that 'our results address the kinetic side of this coherence-growth problem' could be better qualified, since the simulations are 2D pair-plasma with a guide field and no expansion/driving.
  5. [References] The manuscript cites 'Z. Liu et al. 2025a' and 'H. Zhou et al. 2022' among others; please ensure all references are complete and consistently formatted (some entries lack page numbers or DOI).

Circularity Check

0 steps flagged

No significant circularity: the central claims are empirical fits and independent diagnostics, not quantities forced by construction.

full rationale

The paper's central chain is: (i) measure b^2(t) and xi_b(t), (ii) find an empirical energy-scale relation b^2 xi_b^2 ~ const from a plot of b^2 against xi_b (Fig. 2), (iii) independently fit power-law exponents p and q from b^2(t) and xi_b(t) (Fig. 3), and (iv) note p ~ 2q as a consistency check. Section 3 explicitly calls Eq. (17) an 'empirical conservation-like constraint,' not a microscopic invariant, and Section 4 says the constraint 'does not determine the decay exponents p or q ... but only a combined relation p ~ 2q.' Thus p and q are not derived by assuming Eq. (17); they are measured from time series, and p ~ 2q is a posterior consistency statement rather than a construction. The spectral-peak exponent q_peak is an independent diagnostic (Eq. 19), and the broken-self-similarity conclusion is based on time-dependent fitted spectral indices s1, s2, s3, not on an assumed normalization. The self-citations (e.g., H. Zhou et al. 2022, including a coauthor of the present paper, and Hakobyan et al. 2025, also including a coauthor) provide MHD context and code description; they are not load-bearing for the new kinetic result. The main substantive concern--that the exponents are fitted over only 20 <= t/l0 <= 100 with no uncertainties, one realization per sigma_0, and spectral indices still evolving at the end of the simulation--is a legitimate statistical-robustness and asymptotic-convergence concern, but it is not an instance of definitional circularity, fitted input called prediction, or self-citation load-bearing reasoning. No equation in the paper reduces by construction to a fitted value, nor is any central premise justified only by the authors' own prior work.

Axiom & Free-Parameter Ledger

5 free parameters · 4 axioms · 0 invented entities

The paper's central numbers are all fitted outputs from five PIC runs; no new particles, forces, or conserved quantities are introduced. The honest ledger entry is that these fits carry no quoted uncertainties and the physics interpretation depends on standard kinetic/MHD background assumptions rather than new postulated entities.

free parameters (5)
  • Decay exponent p(σ0) = p = 0.58, 0.66, 0.73, 0.77, 0.75 for σ0 = 0.25, 1, 4, 16, 64
    Fitted slope of b²(t) over 20≤t/l0≤100. The claim 'p<1' and its σ0 trend rest directly on these fits. No uncertainties are quoted.
  • Decay exponent q(σ0) = q = 0.27, 0.31, 0.33, 0.38, 0.39
    Fitted slope of ξ_b(t). The claim 'q<1/2' rests on this fit, which is also used to verify p≈2q.
  • Peak-wavenumber exponent q_peak(σ0) = q_peak = 0.43, 0.43, 0.48, 0.56, 0.50
    Fitted slope of k_peak(t). Used as an independent diagnostic of broken self-similarity.
  • Post-peak spectral slopes s1, s2, s3 (median) = s1 ≈ −1.6 to −1.9; s2 ≈ −2.5 to −3.0; s3 ≈ −4.1 to −4.7
    Piecewise-linear/BIC fits used to claim the post-peak spectrum is not a single power law and evolves with time.
  • Island-area tail exponent α = α ≈ 2.05
    Fit to the initial P(S) ∝ S^{−α} distribution, used to support the initial-population contribution to the slow decay.
axioms (4)
  • domain assumption 2.5D pair-plasma PIC at ~1.6 cells per cold skin depth d0 adequately resolves the kinetic physics claimed (Larmor-scale structures, pressure anisotropy).
    Section 2. Convergence is tested only for σ0=16; if resolution is marginal, spectral break locations and fitted exponents could shift.
  • domain assumption CGL double-adiabatic invariants (Chew et al. 1956) govern the pressure-anisotropy evolution.
    Section 6.2, Eq. (31). Assumes no pitch-angle scattering, an approximation for pair plasmas; used to interpret p⊥/p∥ evolution.
  • domain assumption The MHD island-merger reference predictions (p=1, q=1/2) assume equal-sized islands, incompressibility, and a generation-independent reconnection rate.
    Section 1, Eqs. (1)–(5). Used as the benchmark; the paper itself notes incompressibility may not hold in collisionless plasmas (Section 3).
  • standard math Single-scale self-similarity implies E_b(k,t)=b²ξ_b Φ(kξ_b).
    Eq. (21) after Olesen (1997). This is the null hypothesis the paper tests and rejects through spectral-collapse and break analysis.

reviewed 2026-08-02 · how reviews work

0 comments
Cite this review

Pith. "Pith review of Inverse Transfer in Non-helical 2D Collisionless Magnetic Turbulence: Island-Merger Picture with Kinetic Effects." pith.science (2026). https://pith.science/paper/7XBPPCSO

@misc{pith2026260713406,
  author       = {Pith},
  title        = {Pith review of: Inverse Transfer in Non-helical 2D Collisionless Magnetic Turbulence: Island-Merger Picture with Kinetic Effects},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/7XBPPCSO}},
  note         = {Machine review of arXiv:2607.13406}
}
Share X Bluesky LinkedIn Reddit HN
abstract

Magnetic inverse transfer is often invoked to connect small-scale magnetic-field generation to larger coherence scales in high-energy and cosmological plasmas. The underlying magnetohydrodynamic (MHD) arguments combine two logically distinct ingredients: a bulk quantity that is asymptotically conserved in the limit of small resistivity, and a time scale determined by the decay dynamics. In this work, we explore whether this scenario still holds in decaying nonhelical turbulence formed by collisionless plasmas using particle-in-cell simulations. The simulations approximately satisfy $B^2\xi_B^2\simeq{\rm const}$ as in the MHD case, and the fitted exponents in $B^2\propto t^{-p}$ and $\xi_B\propto t^q$ obey $p\simeq2q$. Here $B^2\equiv\langle B_x^2+B_y^2\rangle$ is the average in-plane magnetic energy density, and $\xi_b$ is the magnetic integral scale. However, the decay time scale differs from the MHD case as inferred from the decay exponents. We found $p<1$ and $q<1/2$ in all cases with different initial magnetization $\sigma_0$, with both exponents lower than the MHD values and varying systematically with $\sigma_0$. The spectral peak also migrates toward lower wavenumber at a rate faster than the growth of $\xi_B$, indicating a broken self-similarity. The broken self-similarity is attributed to the appearance of kinetic scales in the magnetic energy spectrum due to pressure anisotropy and Larmor-scale magnetic structures. These results indicate that in astrophysical collisionless plasmas, including but not restrict to solar wind, pulsar-wind nebulae, interstellar medium, and cosmological plasmas, magnetic coherence can continue to grow by inverse transfer, but extrapolations based on MHD decay-time scaling can overestimate the rate of large-scale field growth.

Figures

Figures reproduced from arXiv: 2607.13406 by Hongzhe Zhou, Yangyang Cai, Yosuke Mizuno.

Figure 1
Figure 1. Figure 1: Evolution of a representative σ0 = 1 simulation. For a clear view of the islands, only a (500d0) 2 (Full length L = 5120d0) area is presented. Colors show the total number density on a linear scale, and white contours indicate the in-plane magnetic flux function Az, reconstructed from bx = ∂yAz and by = −∂xAz. The solid and dashed lines show positive and negative Az, respectively. The three snapshots, at t… view at source ↗
Figure 2
Figure 2. Figure 2: Relation between the in-plane magnetic-energy proxy b 2 and the magnetic integral scale ξb for all five magne￾tizations, starting each case at the first sample with b 2 ≃ 0.2. For reference, the red dashed line has slope −2.0. The ap￾proximate alignment with this slope indicates b 2 ξ 2 b ≃ const. time is proportional to the local merger time, giving q = 1/2 and p = 1 (M. Zhou et al. 2019) [PITH_FULL_IMAG… view at source ↗
Figure 3
Figure 3. Figure 3: Power-law fits for b 2 (t) and ξb(t). Colored solid curves show the simulation data, same-color dashed lines show the power-law fits, and the shaded region marks the fitted interval 20 ≤ t/l0 ≤ 100. Text labels give the fitted exponents [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: The exponents shown in this figure are defined by b 2 ∼ t −p , ξb ∼ t q , and kpeak ∼ t −qpeak . The purple line represents the ratio q/p, confirming the conservation law. The black dotted and black dashed lines mark 0.5 and 1.0, respectively, for reference. ond fitted spectral breaks, respectively. The first break lies near kb1ρL ∼ 0.5–0.9, and the second one lies near kb2ρL ∼ 1.0–2.4, indicating that Lar… view at source ↗
Figure 5
Figure 5. Figure 5: Representative magnetic-energy spectra for σ0 = 1 at early (t/l0 ≃ 19.9) and late (t/l0 ≃ 124.9) times. Dashed lines show three-piece log-log fits to the post-peak spectrum over the resolved range kd0 ≤ 1. Dash-dotted vertical lines mark kρL = 1 for the two displayed times, where ρL is the averaged Larmor radius of particles. The listed indices show that the post-peak spectral shape evolves during the deca… view at source ↗
Figure 6
Figure 6. Figure 6: Self-similar spectrum-collapse test. Spectra are plotted as Eb(k, t)/(b 2 ξb) versus kξb for the main dynamical interval. The energy-containing range collapses well, while the low- and high-kξb tails retain systematic residuals [PITH_FULL_IMAGE:figures/full_fig_p008_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Log-time evolution of the fitted post-peak spectral indices. The fit uses kd0 ≤ 1. The three panels show the lower-k post-peak segment s1, the transition segment s2, and the highest-k segment s3. the effective Alf´ven velocity for our system is vA,eff = b q 4πh + b 2 + b 2 g √ T . (28) The effect of pressure aniostropy is discussed in the fol￾lowing subsection. Here we focus on the denominator of Equation … view at source ↗
Figure 8
Figure 8. Figure 8: Logarithmic contribution of the denominator 4πh + b 2 + b 2 g in Equation (28) to the Alfv´en-speed scaling. The plotted quantity is (1/2)d ln(4πh + b 2 + b 2 g)/d ln t over the main dynamical interval. pose two possible origins of the slow decay. The first is an initial-population effect: our initialization does not create identical islands. A finite-width island ensemble can already produce a somewhat sl… view at source ↗
Figure 9
Figure 9. Figure 9: Initial magnetic-island distribution. Left: signed flux normalized by Brms,0l0. The two polarity populations form two approximately Gaussian peaks, shown by the black dashed curves. Right: island area normalized by l 2 0, with a robust S −2 tail. parallel motion has a longitudinal adiabatic invariant. In the form by G. F. Chew et al. (1956), these give p⊥ nBtot ≃ const, p∥B2 tot n3 ≃ const. (31) where Btot… view at source ↗
Figure 10
Figure 10. Figure 10: Pressure-anisotropy phase plots for the five magnetizations. Each panel shows the domain-averaged trajectory in (β∥, p⊥/p∥), colored by time. Each contour encloses 90% of the plasma-cell probability distribution at one sampled time, with the outer 10% lying outside the boundary. The dashed curves mark ∆β∥/2 = ±1, where ∆ = p⊥/p∥ − 1; the lower branch corresponds to a strong reduction of the effective magn… view at source ↗
Figure 11
Figure 11. Figure 11: Log-time evolution of the rms effective-tension diagnostic Trms = (vA,eff,rms/vA,0,rms) 2 . The shaded region marks the main dynamical interval 20 ≤ t/l0 ≤ 100, and the dotted line marks Trms = 1. Lower magnetization runs show a stronger reduction of the effective tension. may mark the onset of an anisotropy-driven flattened component rather than the ordinary sub-Larmor cas￾cade alone. Pressure anisotropy… 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

36 extracted references · 10 canonical work pages

  1. [1]

    2004, Physical Review D, 70, 123003, doi: 10.1103/PhysRevD.70.123003

    Banerjee, R., & Jedamzik, K. 2004, Physical Review D, 70, 123003, doi: 10.1103/PhysRevD.70.123003

    doi:10.1103/physrevd.70.123003 2004

  2. [2]

    Bhat, P., Zhou, M., & Loureiro, N. F. 2021, Monthly Notices of the Royal Astronomical Society, 501, 3074, doi: 10.1093/mnras/staa3849 Kinetic Scaling of Magnetic Inverse Transfer13

    doi:10.1093/mnras/staa3849 2021

  3. [3]

    2017, Physical Review Letters, 118, 055102, doi: 10.1103/PhysRevLett.118.055102

    Brandenburg, A., & Kahniashvili, T. 2017, Physical Review Letters, 118, 055102, doi: 10.1103/PhysRevLett.118.055102

    work page doi:10.1103/physrevlett.118.055102 2017

  4. [4]

    Brandenburg, A., Kahniashvili, T., & Tevzadze, A. G. 2015, Physical Review Letters, 114, 075001, doi: 10.1103/PhysRevLett.114.075001

    doi:10.1103/physrevlett.114.075001 2015

  5. [5]

    2007, Compressible Turbulence in Galaxy Clusters: Physics and Stochastic Particle Re-acceleration,, arXiv preprint astro-ph/0703591 https://arxiv.org/abs/astro-ph/0703591

    Brunetti, G., & Lazarian, A. 2007, Compressible Turbulence in Galaxy Clusters: Physics and Stochastic Particle Re-acceleration,, arXiv preprint astro-ph/0703591 https://arxiv.org/abs/astro-ph/0703591

    Pith/arXiv arXiv 2007

  6. [6]

    2005, Living Reviews in Solar Physics, 2, 4, doi: 10.12942/lrsp-2005-4

    Bruno, R., & Carbone, V. 2005, Living Reviews in Solar Physics, 2, 4, doi: 10.12942/lrsp-2005-4

    work page doi:10.12942/lrsp-2005-4 2005

  7. [7]

    F., Goldberger, M

    Chew, G. F., Goldberger, M. L., & Low, F. E. 1956, Proceedings of the Royal Society of London A, 236, 112, doi: 10.1098/rspa.1956.0116

    arXiv 1956

  8. [8]

    2001, Physical Review E, 64, 056405, doi: 10.1103/PhysRevE.64.056405

    Christensson, M., Hindmarsh, M., & Brandenburg, A. 2001, Physical Review E, 64, 056405, doi: 10.1103/PhysRevE.64.056405

    doi:10.1103/physreve.64.056405 2001

  9. [9]

    2018, Physical Review Letters, 121, 255101, doi: 10.1103/PhysRevLett.121.255101

    Comisso, L., & Sironi, L. 2018, Physical Review Letters, 121, 255101, doi: 10.1103/PhysRevLett.121.255101

    doi:10.1103/physrevlett.121.255101 2018

  10. [10]

    2013, Astronomy and Astrophysics Review, 21, 62, doi: 10.1007/s00159-013-0062-7

    Durrer, R., & Neronov, A. 2013, Astronomy and Astrophysics Review, 21, 62, doi: 10.1007/s00159-013-0062-7

    doi:10.1007/s00159-013-0062-7 2013

  11. [11]

    1993, Physical Review E, 47, 4354, doi: 10.1103/PhysRevE.47.4354

    Gedalin, M. 1993, Physical Review E, 47, 4354, doi: 10.1103/PhysRevE.47.4354

    doi:10.1103/physreve.47.4354 1993

  12. [12]

    M., Cai, Y., et al

    Hakobyan, H., B¨ oss, L. M., Cai, Y., et al. 2025, Entity – Hardware-Agnostic Particle-in-Cell Code for Plasma Astrophysics. I: Curvilinear Special Relativistic Module,, arXiv preprint arXiv:2511.17710 doi: 10.48550/arXiv.2511.17710

    doi:10.48550/arxiv.2511.17710 2025

  13. [13]

    2017, Mirror instability in the turbulent solar wind,, arXiv preprint arXiv:1703.07377 https://arxiv.org/abs/1703.07377

    Hellinger, P., Landi, S., Matteini, L., Verdini, A., & Franci, L. 2017, Mirror instability in the turbulent solar wind,, arXiv preprint arXiv:1703.07377 https://arxiv.org/abs/1703.07377

    Pith/arXiv arXiv 2017

  14. [14]

    2015, Plasma turbulence and kinetic instabilities at ion scales in the expanding solar wind,, arXiv preprint arXiv:1508.03159 https://arxiv.org/abs/1508.03159

    Hellinger, P., Matteini, L., Landi, S., et al. 2015, Plasma turbulence and kinetic instabilities at ion scales in the expanding solar wind,, arXiv preprint arXiv:1508.03159 https://arxiv.org/abs/1508.03159

    Pith/arXiv arXiv 2015

  15. [15]

    N., & Schekochihin, A

    Hosking, D. N., & Schekochihin, A. A. 2021, Physical Review X, 11, 041005, doi: 10.1103/PhysRevX.11.041005

    doi:10.1103/physrevx.11.041005 2021

  16. [16]

    W., Schekochihin, A

    Kunz, M. W., Schekochihin, A. A., & Stone, J. M. 2014, Physical Review Letters, 112, 205003, doi: 10.1103/PhysRevLett.112.205003

    doi:10.1103/physrevlett.112.205003 2014

  17. [17]

    2013, Monthly Notices of the Royal Astronomical Society, 428, 845, doi: 10.1093/mnras/sts081

    Lemoine, M. 2013, Monthly Notices of the Royal Astronomical Society, 428, 845, doi: 10.1093/mnras/sts081

    doi:10.1093/mnras/sts081 2013

  18. [18]

    Liu, Z., Zhou, M., & Loureiro, N. F. G. 2025a, The Astrophysical Journal Letters, 995, L46, doi: 10.3847/2041-8213/ae247e

    work page doi:10.3847/2041-8213/ae247e 2041

  19. [19]

    2005, Physical review letters, 95, 235003

    Schekochihin, A. 2005, Physical review letters, 95, 235003

    2005

  20. [20]

    F., Schekochihin, A

    Loureiro, N. F., Schekochihin, A. A., & Cowley, S. C. 2007, Physics of Plasmas, 14, 100703, doi: 10.1063/1.2783986

    doi:10.1063/1.2783986 2007

  21. [21]

    V., & Loeb, A

    Medvedev, M. V., & Loeb, A. 1999, The Astrophysical Journal, 526, 697, doi: 10.1086/308038

    doi:10.1086/308038 1999

  22. [22]

    A., & Kunz, M

    Melville, S., Schekochihin, A. A., & Kunz, M. W. 2016, Monthly Notices of the Royal Astronomical Society, 459, 2701, doi: 10.1093/mnras/stw793

    doi:10.1093/mnras/stw793 2016

  23. [23]

    1997, Physics Letters B, 398, 321, doi: 10.1016/S0370-2693(97)00235-9

    Olesen, P. 1997, Physics Letters B, 398, 321, doi: 10.1016/S0370-2693(97)00235-9

    work page doi:10.1016/s0370-2693(97)00235-9 1997

  24. [24]

    2010, Physical Review Letters, 105, 131101, doi: 10.1103/PhysRevLett.105.131101

    Rezeau, L. 2010, Physical Review Letters, 105, 131101, doi: 10.1103/PhysRevLett.105.131101

    doi:10.1103/physrevlett.105.131101 2010

  25. [25]

    L., Robert, P., & Khotyaintsev, Y

    Sahraoui, F., Goldstein, M. L., Robert, P., & Khotyaintsev, Y. V. 2009, Physical Review Letters, 102, 231102, doi: 10.1103/PhysRevLett.102.231102

    doi:10.1103/physrevlett.102.231102 2009

  26. [26]

    A., Cowley, S

    Schekochihin, A. A., Cowley, S. C., Kulsrud, R. M., Rosin, M. S., & Heinemann, T. 2008, Physical Review Letters, 100, 081301, doi: 10.1103/PhysRevLett.100.081301

    work page doi:10.1103/physrevlett.100.081301 2008

  27. [27]

    Son, D. T. 1999, Physical Review D, 59, 063008, doi: 10.1103/PhysRevD.59.063008

    work page doi:10.1103/physrevd.59.063008 1999

  28. [28]

    W., Quataert, E., & Schekochihin, A

    Squire, J., Kunz, M. W., Quataert, E., & Schekochihin, A. A. 2017, Physical Review Letters, 119, 155101, doi: 10.1103/PhysRevLett.119.155101

    work page doi:10.1103/physrevlett.119.155101 2017

  29. [29]

    2016, Reports on Progress in Physics, 79, 076901, doi: 10.1088/0034-4885/79/7/076901

    Subramanian, K. 2016, Reports on Progress in Physics, 79, 076901, doi: 10.1088/0034-4885/79/7/076901

    doi:10.1088/0034-4885/79/7/076901 2016

  30. [30]

    2011, Nonlinear Processes in Geophysics, 18, 587, doi: 10.5194/npg-18-587-2011

    Voitenko, Y., & De Keyser, J. 2011, Nonlinear Processes in Geophysics, 18, 587, doi: 10.5194/npg-18-587-2011

    work page doi:10.5194/npg-18-587-2011 2011

  31. [31]

    Begelman, M. C. 2018, Monthly Notices of the Royal Astronomical Society, 474, 2514, doi: 10.1093/mnras/stx2883

    doi:10.1093/mnras/stx2883 2018

  32. [32]

    Begelman, M. C. 2017, Physical Review Letters, 118, 055103, doi: 10.1103/PhysRevLett.118.055103

    doi:10.1103/physrevlett.118.055103 2017

  33. [33]

    2022, Journal of Plasma Physics, 88, 905880602, doi: 10.1017/S002237782200109X

    Zhou, H., Sharma, R., & Brandenburg, A. 2022, Journal of Plasma Physics, 88, 905880602, doi: 10.1017/S002237782200109X

    doi:10.1017/s002237782200109x 2022

  34. [34]

    F., & Uzdensky, D

    Zhou, M., Bhat, P., Loureiro, N. F., & Uzdensky, D. A. 2019, Physical Review Research, 1, 012004, doi: 10.1103/PhysRevResearch.1.012004

    doi:10.1103/physrevresearch.1.012004 2019

  35. [35]

    H., & Loureiro, N

    Zhou, M., Wu, D. H., & Loureiro, N. F. 2021, Journal of Plasma Physics, 87, 905870602, doi: 10.1017/S0022377821001112

    work page doi:10.1017/s0022377821001112 2021

  36. [36]

    2014, The Astrophysical Journal Letters, 794, L26, doi: 10.1088/2041-8205/794/2/L26

    Zrake, J. 2014, The Astrophysical Journal Letters, 794, L26, doi: 10.1088/2041-8205/794/2/L26

    work page doi:10.1088/2041-8205/794/2/l26 2014

This paper was first reviewed by deepseek-v4-flash on August 2, 2026.