Gyrokinetic-MHD Theory in Fusion Plasmas

• Gyrokinetic-MHD theory is a unified framework that blends kinetic gyrokinetic effects with fluid MHD dynamics to capture low-frequency, long-wavelength plasma behavior.
• It incorporates key kinetic features such as finite Larmor radius effects, electron inertia, and collisionless dissipation to accurately model instabilities like internal kinks, tearing modes, and Alfvén eigenmodes.
• Hybrid numerical implementations and rigorous field-theoretic foundations ensure energetic consistency and enable practical applications in predicting tokamak disruptions and optimizing fusion performance.

Gyrokinetic-MHD theory is a unified framework integrating gyrokinetic and magnetohydrodynamic (MHD) descriptions to capture both kinetic and fluid aspects of low-frequency, long-wavelength collective dynamics in strongly magnetized plasmas. It subsumes the ideal MHD limit as a particular case in the long-wavelength, low-$\beta$, isotropic-pressure regime, while incorporating essential kinetic effects such as finite Larmor radius (FLR), electron inertia, and collisionless dissipation. This synthesis underpins predictive models of macroscopic instabilities, Alfvén eigenmodes, and energetic particle physics in toroidal confinement devices.

1. Fundamental Equations and Reduction to MHD

The collisionless gyrokinetic system describes the evolution of the perturbed distribution function $f_{s1}$ for species $s$ in guiding-center phase space, subject to Maxwell's equations for the electrostatic potential $\Phi$ and vector potential components $A_\parallel$: $\begin{align*} &\text{Quasi-neutrality:} && -\nabla\cdot\left[n_{i0}\left(\frac{m_i}{B_0^2}\right)\nabla_\perp\Phi\right] = \sum_{s=i,e}q_s n_{s1} \ &\text{Parallel Ampère:} && \partial_tA_\parallel^{(s)} + b_0\cdot\nabla\Phi = 0 \ &\text{Helmholtz for %%%%5%%%%:} && \left[\sum_{s=i,e}\mu_0 q_s^2 n_{s0}/m_s - \nabla_\perp^2\right]A_\parallel^{(h)} = \mu_0\sum_{s=i,e}j_{s1,\parallel} + \nabla_\perp^2A_\parallel^{(s)} \ &\text{Gyrokinetic Vlasov:} && \partial_t f_{s1} + \dot R\cdot\nabla_R f_{s1} + \dot v_\parallel \partial_{v_\parallel}f_{s1} = -\dot R^{(1)}\cdot\nabla_R f_{s0} - \dot\epsilon^{(1)} \partial_\epsilon f_{s0} \end{align*}$ where $\dot R$ and $\dot v_\parallel$ are the guiding center and parallel phase-space velocities.

Taking velocity-space moments and applying the long-wavelength ($k_\perp\rho_i\ll 1$), low-$\beta$, and isotropy ($f_{s1}$0) ordering, these equations reduce exactly to the single-fluid linearized MHD vorticity balance: $f_{s1}$1 demonstrating that collisionless gyrokinetics inherently contains the ideal MHD limit (Antlitz et al., 14 Oct 2025).

2. Linear Instability Physics and Dispersion Relations

Internal Kink Modes

The classical $f_{s1}$2 internal kink instability at the $f_{s1}$3 surface is governed by the radial displacement eigenmode: $f_{s1}$4 with parameters $f_{s1}$5 and $f_{s1}$6 poloidal field and geometry factors. The growth rate is given by $f_{s1}$7 and the structure of the mode strongly depends on inclusion of diamagnetic effects, parallel field fluctuations $f_{s1}$8, the treatment of FLR, and mass ratio effects (Antlitz et al., 14 Oct 2025).

Collisionless Tearing Modes

Electron inertia enables collisionless tearing ($f_{s1}$9) at the $s$0 surface, inaccessible in pure MHD. The growth rates scale as:

• $s$1 for $s$2 ($s$3)
• $s$4 for $s$5 ($s$6) where $s$7 is the electron skin depth and $s$8 the Alfvén frequency.

Gyrokinetic eigenvalue solvers confirm these kinetic scaling laws in tearing regime and recover MHD results in the ideal limit (Antlitz et al., 14 Oct 2025, Bao et al., 2017).

Diamagnetic Stabilization

Diamagnetic drift frequencies $s$9 enter via a quadratic equation for normal mode frequency: $\Phi$0 yielding Doppler-shifted, stabilized, or marginally stable branches as $\Phi$1 varies relative to the ideal MHD growth rate. Finite $\Phi$2 always reduces the growth rate and induces a mode frequency $\Phi$3 (Antlitz et al., 14 Oct 2025).

3. Role of Key Parameters and Model Features

Inclusion or neglect of certain physical elements produces significant quantitative and qualitative changes:

• Parallel magnetic perturbation ($\Phi$4): Inclusion is essential; omission gives spurious stabilization or incorrect mode structure (Antlitz et al., 14 Oct 2025, Wei et al., 2021).
• Aspect ratio ($\Phi$5): The kink mode stabilizes for small $\Phi$6; reduced-MHD models become inaccurate at low aspect ratio (Antlitz et al., 14 Oct 2025).
• Ion model (drift vs gyrokinetic): Only fully gyrokinetic ions (with FLR) recover correct frequency scaling and stabilization. Drift-kinetic treatments miss FLR stabilization and underestimate $\Phi$7 (Antlitz et al., 14 Oct 2025).
• Electron-to-ion mass ratio ($\Phi$8): Critical for capturing collisionless tearing mode scaling; MHD limit ($\Phi$9) misses this behavior (Antlitz et al., 14 Oct 2025).
• Magnetic compressibility: In gyrokinetic field theory, the cancellation of compressional and grad-B drift currents at order $A_\parallel$0 is essential to energetically consistent dynamics and recovers the cancellation found in kinetic-ballooning mode theory (Scott, 2024).

Model Feature	Ideal MHD	Extended MHD ($A_\parallel$1)	Gyrokinetic ($A_\parallel$2)	Gyrokinetic (no $A_\parallel$3)
1/1 kink growth $A_\parallel$4	$A_\parallel$5	Reduced by $A_\parallel$6	Matches extended MHD	Underestimates $A_\parallel$7
1/1 kink frequency $A_\parallel$8	0	$A_\parallel$9	$\begin{align*} &\text{Quasi-neutrality:} && -\nabla\cdot\left[n_{i0}\left(\frac{m_i}{B_0^2}\right)\nabla_\perp\Phi\right] = \sum_{s=i,e}q_s n_{s1} \ &\text{Parallel Ampère:} && \partial_tA_\parallel^{(s)} + b_0\cdot\nabla\Phi = 0 \ &\text{Helmholtz for %%%%5%%%%:} && \left[\sum_{s=i,e}\mu_0 q_s^2 n_{s0}/m_s - \nabla_\perp^2\right]A_\parallel^{(h)} = \mu_0\sum_{s=i,e}j_{s1,\parallel} + \nabla_\perp^2A_\parallel^{(s)} \ &\text{Gyrokinetic Vlasov:} && \partial_t f_{s1} + \dot R\cdot\nabla_R f_{s1} + \dot v_\parallel \partial_{v_\parallel}f_{s1} = -\dot R^{(1)}\cdot\nabla_R f_{s0} - \dot\epsilon^{(1)} \partial_\epsilon f_{s0} \end{align*}$0 ($\begin{align*} &\text{Quasi-neutrality:} && -\nabla\cdot\left[n_{i0}\left(\frac{m_i}{B_0^2}\right)\nabla_\perp\Phi\right] = \sum_{s=i,e}q_s n_{s1} \ &\text{Parallel Ampère:} && \partial_tA_\parallel^{(s)} + b_0\cdot\nabla\Phi = 0 \ &\text{Helmholtz for %%%%5%%%%:} && \left[\sum_{s=i,e}\mu_0 q_s^2 n_{s0}/m_s - \nabla_\perp^2\right]A_\parallel^{(h)} = \mu_0\sum_{s=i,e}j_{s1,\parallel} + \nabla_\perp^2A_\parallel^{(s)} \ &\text{Gyrokinetic Vlasov:} && \partial_t f_{s1} + \dot R\cdot\nabla_R f_{s1} + \dot v_\parallel \partial_{v_\parallel}f_{s1} = -\dot R^{(1)}\cdot\nabla_R f_{s0} - \dot\epsilon^{(1)} \partial_\epsilon f_{s0} \end{align*}$1)	$\begin{align*} &\text{Quasi-neutrality:} && -\nabla\cdot\left[n_{i0}\left(\frac{m_i}{B_0^2}\right)\nabla_\perp\Phi\right] = \sum_{s=i,e}q_s n_{s1} \ &\text{Parallel Ampère:} && \partial_tA_\parallel^{(s)} + b_0\cdot\nabla\Phi = 0 \ &\text{Helmholtz for %%%%5%%%%:} && \left[\sum_{s=i,e}\mu_0 q_s^2 n_{s0}/m_s - \nabla_\perp^2\right]A_\parallel^{(h)} = \mu_0\sum_{s=i,e}j_{s1,\parallel} + \nabla_\perp^2A_\parallel^{(s)} \ &\text{Gyrokinetic Vlasov:} && \partial_t f_{s1} + \dot R\cdot\nabla_R f_{s1} + \dot v_\parallel \partial_{v_\parallel}f_{s1} = -\dot R^{(1)}\cdot\nabla_R f_{s0} - \dot\epsilon^{(1)} \partial_\epsilon f_{s0} \end{align*}$2
Collisionless tearing	No branch	No branch	Captured ($\begin{align*} &\text{Quasi-neutrality:} && -\nabla\cdot\left[n_{i0}\left(\frac{m_i}{B_0^2}\right)\nabla_\perp\Phi\right] = \sum_{s=i,e}q_s n_{s1} \ &\text{Parallel Ampère:} && \partial_tA_\parallel^{(s)} + b_0\cdot\nabla\Phi = 0 \ &\text{Helmholtz for %%%%5%%%%:} && \left[\sum_{s=i,e}\mu_0 q_s^2 n_{s0}/m_s - \nabla_\perp^2\right]A_\parallel^{(h)} = \mu_0\sum_{s=i,e}j_{s1,\parallel} + \nabla_\perp^2A_\parallel^{(s)} \ &\text{Gyrokinetic Vlasov:} && \partial_t f_{s1} + \dot R\cdot\nabla_R f_{s1} + \dot v_\parallel \partial_{v_\parallel}f_{s1} = -\dot R^{(1)}\cdot\nabla_R f_{s0} - \dot\epsilon^{(1)} \partial_\epsilon f_{s0} \end{align*}$3)	Absent

4. Hybrid Kinetic-MHD Models and Numerical Implementations

Hybrid frameworks, such as the GMEC and XHMGC codes, combine reduced MHD solvers for bulk plasma variables with gyrokinetic particle simulation for energetic or thermal ions. The generic coupling is:

$\begin{align*} &\text{Quasi-neutrality:} && -\nabla\cdot\left[n_{i0}\left(\frac{m_i}{B_0^2}\right)\nabla_\perp\Phi\right] = \sum_{s=i,e}q_s n_{s1} \ &\text{Parallel Ampère:} && \partial_tA_\parallel^{(s)} + b_0\cdot\nabla\Phi = 0 \ &\text{Helmholtz for %%%%5%%%%:} && \left[\sum_{s=i,e}\mu_0 q_s^2 n_{s0}/m_s - \nabla_\perp^2\right]A_\parallel^{(h)} = \mu_0\sum_{s=i,e}j_{s1,\parallel} + \nabla_\perp^2A_\parallel^{(s)} \ &\text{Gyrokinetic Vlasov:} && \partial_t f_{s1} + \dot R\cdot\nabla_R f_{s1} + \dot v_\parallel \partial_{v_\parallel}f_{s1} = -\dot R^{(1)}\cdot\nabla_R f_{s0} - \dot\epsilon^{(1)} \partial_\epsilon f_{s0} \end{align*}$4

where $\begin{align*} &\text{Quasi-neutrality:} && -\nabla\cdot\left[n_{i0}\left(\frac{m_i}{B_0^2}\right)\nabla_\perp\Phi\right] = \sum_{s=i,e}q_s n_{s1} \ &\text{Parallel Ampère:} && \partial_tA_\parallel^{(s)} + b_0\cdot\nabla\Phi = 0 \ &\text{Helmholtz for %%%%5%%%%:} && \left[\sum_{s=i,e}\mu_0 q_s^2 n_{s0}/m_s - \nabla_\perp^2\right]A_\parallel^{(h)} = \mu_0\sum_{s=i,e}j_{s1,\parallel} + \nabla_\perp^2A_\parallel^{(s)} \ &\text{Gyrokinetic Vlasov:} && \partial_t f_{s1} + \dot R\cdot\nabla_R f_{s1} + \dot v_\parallel \partial_{v_\parallel}f_{s1} = -\dot R^{(1)}\cdot\nabla_R f_{s0} - \dot\epsilon^{(1)} \partial_\epsilon f_{s0} \end{align*}$5 is the non-adiabatic energetic particle pressure returning from the gyrokinetic PIC subsystem (Liu et al., 2024, Wang et al., 2010).

The evolution of energetic particle markers is: $\begin{align*} &\text{Quasi-neutrality:} && -\nabla\cdot\left[n_{i0}\left(\frac{m_i}{B_0^2}\right)\nabla_\perp\Phi\right] = \sum_{s=i,e}q_s n_{s1} \ &\text{Parallel Ampère:} && \partial_tA_\parallel^{(s)} + b_0\cdot\nabla\Phi = 0 \ &\text{Helmholtz for %%%%5%%%%:} && \left[\sum_{s=i,e}\mu_0 q_s^2 n_{s0}/m_s - \nabla_\perp^2\right]A_\parallel^{(h)} = \mu_0\sum_{s=i,e}j_{s1,\parallel} + \nabla_\perp^2A_\parallel^{(s)} \ &\text{Gyrokinetic Vlasov:} && \partial_t f_{s1} + \dot R\cdot\nabla_R f_{s1} + \dot v_\parallel \partial_{v_\parallel}f_{s1} = -\dot R^{(1)}\cdot\nabla_R f_{s0} - \dot\epsilon^{(1)} \partial_\epsilon f_{s0} \end{align*}$6 with scattered pressures re-coupled into the MHD equations as moment closures.

High-order finite-volume and field-aligned coordinate systems are standard for accurate global toroidal simulations of Alfvénic modes, kinks, tearing, and energetic particle-driven instabilities (Liu et al., 2024).

5. Field-Theoretic and Hamiltonian Foundations

Gyrokinetic-MHD models are fundamentally derived from variational field theories: $\begin{align*} &\text{Quasi-neutrality:} && -\nabla\cdot\left[n_{i0}\left(\frac{m_i}{B_0^2}\right)\nabla_\perp\Phi\right] = \sum_{s=i,e}q_s n_{s1} \ &\text{Parallel Ampère:} && \partial_tA_\parallel^{(s)} + b_0\cdot\nabla\Phi = 0 \ &\text{Helmholtz for %%%%5%%%%:} && \left[\sum_{s=i,e}\mu_0 q_s^2 n_{s0}/m_s - \nabla_\perp^2\right]A_\parallel^{(h)} = \mu_0\sum_{s=i,e}j_{s1,\parallel} + \nabla_\perp^2A_\parallel^{(s)} \ &\text{Gyrokinetic Vlasov:} && \partial_t f_{s1} + \dot R\cdot\nabla_R f_{s1} + \dot v_\parallel \partial_{v_\parallel}f_{s1} = -\dot R^{(1)}\cdot\nabla_R f_{s0} - \dot\epsilon^{(1)} \partial_\epsilon f_{s0} \end{align*}$7 with single-particle Lagrangians

$\begin{align*} &\text{Quasi-neutrality:} && -\nabla\cdot\left[n_{i0}\left(\frac{m_i}{B_0^2}\right)\nabla_\perp\Phi\right] = \sum_{s=i,e}q_s n_{s1} \ &\text{Parallel Ampère:} && \partial_tA_\parallel^{(s)} + b_0\cdot\nabla\Phi = 0 \ &\text{Helmholtz for %%%%5%%%%:} && \left[\sum_{s=i,e}\mu_0 q_s^2 n_{s0}/m_s - \nabla_\perp^2\right]A_\parallel^{(h)} = \mu_0\sum_{s=i,e}j_{s1,\parallel} + \nabla_\perp^2A_\parallel^{(s)} \ &\text{Gyrokinetic Vlasov:} && \partial_t f_{s1} + \dot R\cdot\nabla_R f_{s1} + \dot v_\parallel \partial_{v_\parallel}f_{s1} = -\dot R^{(1)}\cdot\nabla_R f_{s0} - \dot\epsilon^{(1)} \partial_\epsilon f_{s0} \end{align*}$8

Variation yields gyrokinetic Vlasov, Poisson, and Ampère equations, with Noether's theorem guaranteeing exact conservation of total energy and toroidal canonical momentum (Scott et al., 2010, Scott, 2024, Scott, 2017, deOliveira-Lopes et al., 2023).

In the MHD limit (long-wavelength, low-$\begin{align*} &\text{Quasi-neutrality:} && -\nabla\cdot\left[n_{i0}\left(\frac{m_i}{B_0^2}\right)\nabla_\perp\Phi\right] = \sum_{s=i,e}q_s n_{s1} \ &\text{Parallel Ampère:} && \partial_tA_\parallel^{(s)} + b_0\cdot\nabla\Phi = 0 \ &\text{Helmholtz for %%%%5%%%%:} && \left[\sum_{s=i,e}\mu_0 q_s^2 n_{s0}/m_s - \nabla_\perp^2\right]A_\parallel^{(h)} = \mu_0\sum_{s=i,e}j_{s1,\parallel} + \nabla_\perp^2A_\parallel^{(s)} \ &\text{Gyrokinetic Vlasov:} && \partial_t f_{s1} + \dot R\cdot\nabla_R f_{s1} + \dot v_\parallel \partial_{v_\parallel}f_{s1} = -\dot R^{(1)}\cdot\nabla_R f_{s0} - \dot\epsilon^{(1)} \partial_\epsilon f_{s0} \end{align*}$9), these Euler-Lagrange equations reduce precisely to the single-fluid vorticity, momentum, and energy transport equations of ideal MHD, ensuring energetic and dynamical consistency (Scott et al., 2010, Scott, 2017).

6. Physical Interpretation and Unified Theoretical Picture

Gyrokinetic-MHD theory establishes that:

• Gyrokinetics contains MHD: It subsumes single-fluid ideal MHD as a limiting case, with all fluid equations derivable as moment-reductions under appropriate orderings (Antlitz et al., 14 Oct 2025).
• Kinetic generalization: Finite $\dot R$0, electron inertia, FLR, and kinetic closures (Landau, diamagnetic, phase-mixing) extend the model to describe additional branches: collisionless tearing, kinetic-Ballooning, kinetic-Beta-induced Alfvén Eigenmodes (KBAE), and finite-orbit-width effects absent in MHD (Wang et al., 2010, Liu et al., 2024).
• Energetic consistency and Hamiltonian closure: The field-theoretic derivation ensures conservation laws, abelian gauge invariance, and energetic consistency across kinetic–fluid couplings (Scott et al., 2010, deOliveira-Lopes et al., 2023, Scott, 2017).
• Practical implementation: Hybrid and fully kinetic codes benchmarked against analytic theory and MHD codes confirm that careful implementation (e.g., handling of $\dot R$1, equilibrium current $\dot R$2, field-aligned coordinates) is essential for quantitative agreement, especially for core instabilities in tokamaks (Antlitz et al., 14 Oct 2025, Wei et al., 2021, Liu et al., 2024).

7. Outlook and Future Directions

Unified fluid-kinetic models based on gyrokinetic field theory are advancing toward capturing all critical tokamak MHD and kinetic processes—including ideal and resistive instabilities, energetic particle physics, and nonlinear saturation—in a single framework. Next-generation implementations are focusing on seamless coupling of gyrokinetic and fluid closures, improved energetic particle coupling, and fully consistent magnetic geometry representations (e.g., Boozer and straight-field-line coordinates) for realistic simulations of fusion plasma dynamics (Antlitz et al., 14 Oct 2025, Wei et al., 2021, deOliveira-Lopes et al., 2023).

A plausible implication is that such models will become standard in both disruption prediction and burning plasma performance optimization as full-scale computational approaches become more tractable, leveraging the formal connections and energetic consistency established in gyrokinetic-MHD theory.