Tearing-Mode Instability in Plasmas

Updated 14 July 2026

Tearing-mode instability is a reconnecting process in thin current sheets characterized by magnetic island formation and flux unfreezing.
Its growth rate scales with the Lundquist number and sheet aspect ratio, distinguishing classical resistive MHD from ideal and kinetic regimes.
The instability underpins energy release in diverse plasma systems, informing studies in astrophysical contexts, fusion devices, and turbulent reconnection.

Searching arXiv for recent and foundational papers on tearing-mode instability, emphasizing the papers provided and closely related work. Tearing-mode instability is a reconnecting instability of thin current sheets in which a magnetic equilibrium with field reversal becomes unstable to perturbations localized about a resonant surface, so that non-ideal effects break flux freezing and allow the formation of magnetic islands or plasmoids. In the classical picture the instability is organized by the tearing stability parameter $\Delta'$ , the Lundquist number $S$ , and the current-sheet aspect ratio $a/L$ ; in later developments it has been reformulated for “ideal” tearing, Hall and electron-MHD limits, partially ionized and gyrotropic plasmas, gyrokinetic tokamak configurations, and fully three-dimensional current-sheet geometries (Zanna et al., 2016, Landi et al., 2015).

1. Classical resistive-MHD formulation

In resistive MHD, tearing is usually formulated for a one-dimensional current sheet of half-thickness $a$ embedded in a system of length $L$ , with reversing field $B(y)=B_0 F(y/a)\,\hat x$ or, in Harris-sheet form, $B_x(y)=B_0\tanh(y/a)$ . The standard control parameters are the Alfvén speed $V_A=B_0/\sqrt{\mu_0\rho}$ , the Alfvén time $\tau_A=L/V_A$ , the Lundquist number $S\equiv L V_A/\eta$ , and the inverse aspect ratio $S$ 0. In the classical resistive theory, the fastest-growing mode satisfies $S$ 1, while the constant- $S$ 2 approximation for a Harris sheet uses $S$ 3 as the tearing-stability index (Pucci et al., 2020, Shi et al., 2020).

The inner–outer asymptotic structure is central. Outside the resistive layer the perturbation obeys ideal-MHD equations, and inside the narrow reconnecting layer resistive diffusion regularizes the singularity at the resonant surface. Instability requires $S$ 4; in the Harris-sheet normalization this suggests that sufficiently long-wavelength modes satisfy the basic tearing criterion. In local sheet variables, the classical large- $S$ 5 estimate is $S$ 6, where $S$ 7 and $S$ 8 (Zanna et al., 2016).

This classical framework already distinguishes tearing from other current-sheet instabilities. It is driven by magnetic shear and reconnection physics rather than by pure hydrodynamic shear, and it produces O-points and X-points through reconnection of oppositely directed magnetic field lines. That structure remains recognizable across later Hall, kinetic, and toroidal generalizations, even when the inner-layer physics is substantially altered.

2. Ideal tearing and the disruption of thin current sheets

A major reformulation arises when the current-sheet aspect ratio is allowed to scale with $S$ 9 as $a/L$ 0. For the fastest tearing mode, the growth rate then scales as $a/L$ 1. Requiring the exponent to vanish in the $a/L$ 2 limit gives the critical exponent $a/L$ 3, hence the “ideal-tearing” threshold

$a/L$ 4

At that threshold the growth rate becomes independent of $a/L$ 5 to leading order, with $a/L$ 6 and $a/L$ 7; the fastest mode also satisfies $a/L$ 8 and $a/L$ 9 (Landi et al., 2015).

Two-dimensional compressible resistive-MHD simulations for $a$ 0 confirm these scalings. They recover the predicted dispersion curves and eigenmode structure, including the even magnetic perturbation $a$ 1 peaked inside the sheet, the odd transverse flow $a$ 2 vanishing at the origin, and the even parallel flow $a$ 3 in quadrature with $a$ 4 along the ignorable direction. For the least diffusive case reported, $a$ 5, the tearing instability reaches $a$ 6, i.e. macroscopic Alfvénic growth (Landi et al., 2015).

The nonlinear regime is hierarchical. In multi-mode runs, initially formed islands elongate the X-point regions into new local current sheets of half-length $a$ 7 and half-thickness $a$ 8. These secondary sheets satisfy the same critical relation on local scales,

$a$ 9

with measured aspect ratios $L$ 0 at $L$ 1, $L$ 2 at $L$ 3, and $L$ 4 at $L$ 5, fitted by $L$ 6 with $L$ 7– $L$ 8 (Landi et al., 2015). Because $L$ 9, each successive generation reconnects faster than the previous one. This produces a cascading, accelerating plasmoid regime.

The comparison with Sweet–Parker scaling is decisive. A Sweet–Parker sheet has $B(y)=B_0 F(y/a)\,\hat x$ 0, which is much thinner than the critical $B(y)=B_0 F(y/a)\,\hat x$ 1 threshold. The simulations and scaling analysis therefore support the conclusion that a sheet disrupts through tearing before reaching the fully collapsed Sweet–Parker aspect ratio; in that sense, the Sweet–Parker configuration is likely to be never realized in nature for sufficiently large $B(y)=B_0 F(y/a)\,\hat x$ 2 (Zanna et al., 2016).

3. Hall, partial-ionization, and gyrotropic extensions

In electron magnetohydrodynamics, the tearing-type Hall instability is driven by the combined action of Hall-induced whistler oscillations and shear in the electron current velocity. A local necessary condition is that the background field and the vorticity of the electron current velocity be antiparallel:

$B(y)=B_0 F(y/a)\,\hat x$ 3

or equivalently,

$B(y)=B_0 F(y/a)\,\hat x$ 4

For large Hall number $B(y)=B_0 F(y/a)\,\hat x$ 5, asymptotic matching yields $B(y)=B_0 F(y/a)\,\hat x$ 6 for vacuum-wall tearing and $B(y)=B_0 F(y/a)\,\hat x$ 7 for periodic or broad profiles; in physical units, $B(y)=B_0 F(y/a)\,\hat x$ 8 with $B(y)=B_0 F(y/a)\,\hat x$ 9 or $B_x(y)=B_0\tanh(y/a)$ 0 respectively. Perfect-conductor walls strongly suppress or completely quench the mode by forbidding the essential helicoidal motion at the boundaries (Kitchatinov, 2021).

When a guide field and Hall physics are included in resistive Hall-MHD, the fastest linear instability remains the parallel mode with $B_x(y)=B_0\tanh(y/a)$ 1; a strong guide field does not move the global maximum to oblique wavevectors. Instead, it shifts the resonant surface for oblique modes, suppresses their wave-like inner-layer structure, and makes the eigenfunctions asymmetric. In the Hall case, oblique modes acquire a complex frequency and propagate along the guide-field direction, whereas for $B_x(y)=B_0\tanh(y/a)$ 2 the propagation is non-dispersive with $B_x(y)=B_0\tanh(y/a)$ 3 (Shi et al., 2020).

Partial ionization modifies both resistivity and inertia. Electron–neutral collisions enhance the magnetic diffusivity, while ion–neutral collisions change the effective Alfvénic response. In the coupled regime the ideal-tearing threshold becomes

$B_x(y)=B_0\tanh(y/a)$ 4

in the intermediate regime

$B_x(y)=B_0\tanh(y/a)$ 5

and in the decoupled regime the fully ionized result $B_x(y)=B_0\tanh(y/a)$ 6 is recovered up to negligible corrections. The intermediate regime is stabilizing in the sense that ion–neutral coupling acts like an effective viscosity, although the electron–neutral contribution hidden in $B_x(y)=B_0\tanh(y/a)$ 7 can reduce $B_x(y)=B_0\tanh(y/a)$ 8 and thereby shift the net threshold in the opposite direction (Pucci et al., 2020).

Weakly collisional gyrotropic plasmas introduce an additional layer of modification. For a force-free Harris sheet with prescribed equilibrium pressure anisotropy, the classical $B_x(y)=B_0\tanh(y/a)$ 9 scaling of the maximum growth rate is retained but acquires the prefactor $V_A=B_0/\sqrt{\mu_0\rho}$ 0, where $V_A=B_0/\sqrt{\mu_0\rho}$ 1 and $V_A=B_0/\sqrt{\mu_0\rho}$ 2 depends on $V_A=B_0/\sqrt{\mu_0\rho}$ 3, $V_A=B_0/\sqrt{\mu_0\rho}$ 4, and the gyrotropic closure. Positive $V_A=B_0/\sqrt{\mu_0\rho}$ 5 suppresses tearing and broadens the inner layer; negative $V_A=B_0/\sqrt{\mu_0\rho}$ 6 enhances growth and shifts the fastest mode to larger wavenumber (Kowal et al., 21 Jun 2026). A separate non-ideal CGL-MHD treatment with self-consistent pressure-anisotropy fluctuations finds a distinct high- $V_A=B_0/\sqrt{\mu_0\rho}$ 7 scaling even for initially isotropic equilibria,

$V_A=B_0/\sqrt{\mu_0\rho}$ 8

thereby breaking the standard MHD $V_A=B_0/\sqrt{\mu_0\rho}$ 9-independence in the high- $\tau_A=L/V_A$ 0 regime (Ferreira-Santos et al., 16 Mar 2025).

The tearing instability is not confined to strictly two-dimensional sheets. For a flux-tube-like equilibrium of the form $\tau_A=L/V_A$ 1, linear analysis and direct numerical simulations show that a tearing-like mode persists in three dimensions even without a guide field. The principal effect of the modulation $\tau_A=L/V_A$ 2 is a reduction in the net growth rate by the factor

$\tau_A=L/V_A$ 3

while the $\tau_A=L/V_A$ 4- and $\tau_A=L/V_A$ 5-scalings remain similar to the two-dimensional case. The local resistive layer becomes $\tau_A=L/V_A$ 6-dependent, with $\tau_A=L/V_A$ 7, and the simulations show synchronized growth across different $\tau_A=L/V_A$ 8-slices rather than a mere stack of uncoupled two-dimensional sheets (Kumar et al., 2024).

A transverse magnetic-field component can instead quench tearing. In viscoresistive incompressible MHD, introducing $\tau_A=L/V_A$ 9 into an initial Harris sheet destroys the static equilibrium and generates a widening neutral layer with half-width

$S\equiv L V_A/\eta$ 0

The effective sheet width $S\equiv L V_A/\eta$ 1 reduces $S\equiv L V_A/\eta$ 2 and depresses the entire $S\equiv L V_A/\eta$ 3 curve. An approximate stabilization threshold is $S\equiv L V_A/\eta$ 4, and direct simulations confirm that sufficiently strong transverse field produces a broad, stable current layer with strong shear flow instead of a plasmoid chain (Kowal et al., 2024).

The onset problem is also non-modal. For the classical incompressible visco-resistive linear operator, pseudospectral analysis shows strong non-normality and transient amplification even before the asymptotic tearing eigenmode dominates. The maximum possible $S\equiv L V_A/\eta$ 5-norm growth scales as $S\equiv L V_A/\eta$ 6 on times $S\equiv L V_A/\eta$ 7 for $S\equiv L V_A/\eta$ 8, which is faster than the modal tearing timescale $S\equiv L V_A/\eta$ 9. The optimal initial perturbations are localized wave packets concentrated at the current sheet, i.e. pseudomodes rather than eigenmodes (MacTaggart, 2018).

These results complicate a common simplification in which tearing onset is identified solely with the dominant unstable eigenvalue. A plausible implication is that, at large $S$ 00, the early evolution of thin sheets can be controlled by transiently amplified localized noise before the classical exponential tearing branch becomes asymptotically dominant.

5. Kinetic, gyrokinetic, and tokamak manifestations

In toroidal gyrokinetic simulations of classical current-gradient-driven tearing, the instability exhibits a collisionless plateau at low collisionality, a smooth transition to a semi-collisional regime, and a growth rate that follows the Fitzpatrick large- $S$ 01 scaling $S$ 02 rather than the steeper Drake–Lee or classical resistive exponents. Even in the absence of pressure gradients the mode has a residual finite rotation frequency, attributed to toroidal finite-Larmor-radius effects; with pressure gradients it rotates at the electron diamagnetic frequency at low collisionality, while at high collisionality the rotation can reverse to the ion diamagnetic direction (Hornsby et al., 2014).

Microtearing extends tearing parity to ion-Larmor-radius scales. In the Mega Ampere Spherical Tokamak, linear gyrokinetic-Maxwell calculations show a tearing-parity instability with radial width $S$ 03, peak growth at $S$ 04– $S$ 05, and instability for $S$ 06, with maximum growth near $S$ 07. The free energy comes from the electron temperature gradient, but the dominant destabilization is not the classical thermal-force or banana-boundary-layer mechanism; rather, the principal drive is from magnetic drifts, with electrostatic potential coupling also essential (Applegate et al., 2011).

A distinct collisionless fine-scale tearing-parity mode exists even at zero collision frequency in slab gyrokinetics. It is also driven by electron temperature gradient, requires finite electron FLR effects, has $S$ 08, persists in the electrostatic limit, and is actually stabilized by electromagnetic feedback. The analytic reduction identifies it with a tearing-parity slab ETG mode that can be more unstable than the more commonly studied twisting-parity ETG branch (Geng et al., 2020).

Self-consistent collisionless tokamak tearing in global ORB5 simulations adds nonlinear structure. For flat profiles, the linear growth rate obeys $S$ 09 for $S$ 10 and $S$ 11 in the mass-ratio scan; the initial nonlinear island width satisfies $S$ 12. Large islands then drive zonal current redistribution and island-induced $S$ 13 shear flows at the separatrix, which destabilize Kelvin–Helmholtz turbulence and cause strong island decay. At sufficiently high $S$ 14, the classical tearing mode is suppressed and a twisting-parity mode with the same helicity is destabilized instead (Widmer et al., 2024).

In neoclassical tokamak physics, tearing also appears as a seeded nonlinear branch. NIMROD simulations of impurity radiation cooling show a two-step mechanism in which a local helical pressure perturbation drives a diamagnetic-current seed island, after which a heuristic neoclassical electron-viscosity closure generates a perturbed bootstrap current that drives neoclassical tearing-mode growth. In those simulations the nonlinear growth rate is proportional to the electron neoclassical viscosity, while the pure-resistive seed behavior depends on impurity radiation power (Zeng et al., 2024).

6. Observational evidence and astrophysical significance

Direct in-situ evidence for tearing-mode onset prior to reconnection has been reported in the terrestrial magnetotail. Using Cluster electron pitch-angle distributions and neural-network outlier detection, fifteen candidate tearing events were identified, all satisfying the collisionless threshold

$S$ 15

for the longest-wavelength mode. The characteristic sequence is a field-aligned counter-streaming beam phase followed by isotropization and heating, with reconnection signatures occurring minutes later; in the statistical sample, reconnection follows ten of the fifteen events with delays between $S$ 16 min $S$ 17 s and $S$ 18 min $S$ 19 s, with mean $S$ 20 min (Bakrania et al., 2022).

The ideal-tearing framework has immediate astrophysical implications. In the solar corona, the collisional Lundquist number is quoted as $S$ 21, so the critical threshold is $S$ 22 rather than the much thinner Sweet–Parker value. Once a coronal current sheet reaches that critical aspect ratio, tearing and plasmoid formation proceed on Alfvénic times $S$ 23– $S$ 24, providing a route to impulsive flaring. The same hierarchical threshold has been proposed for the magnetotail, pulsar wind nebulae, relativistic pair plasmas, and turbulent current-sheet ensembles in which sheets of many lengths thin until they reach the local $S$ 25 condition (Landi et al., 2015).

In relativistic collisionless pair plasmas, the growth rate depends sensitively on how the current sheet is supported. For pressure-supported sheets with a uniform guide field, hardening the particle spectrum suppresses tearing; for force-free sheets the maximum growth rate becomes independent of spectral hardness. Applied to relativistic MHD turbulence, the same theory yields a tearing-mediated disruption scale for turbulent eddies and a critical magnetization above which charge starvation prevents the tearing instability (Demidov et al., 2024).

Taken together, these developments place tearing-mode instability at the center of contemporary reconnection theory, but not as a monolithic mechanism. In some regimes it is a classical resistive eigenmode; in others it is Hall-mediated, semi-collisional, microtearing, or collisionless and gyrotropic. In two dimensions it readily produces plasmoid chains, whereas in three-dimensional stochastic reconnection it may be important only at the initial stage, with Kelvin–Helmholtz instability later dominating turbulence generation (Kowal et al., 2019). The unifying feature is the conversion of current-sheet free energy into reconnecting magnetic topology once the relevant inner-layer physics allows the ideal constraint to fail.