Plasma Physics
The Tearing Mode Instability: How Current Sheets Break Into Magnetic Islands
Ideal magnetohydrodynamics forbids it: field lines are frozen to the plasma and can never change their topology. Yet the Sun sheds 10²⁵–10²⁶ joules in a single flare, and tokamaks disrupt on millisecond timescales — both because a paper-thin layer of finite resistivity lets magnetic field lines slice apart and reconnect. That process is the tearing mode instability: a sheared, force-balanced current sheet spontaneously fragments into a chain of magnetic islands, converting stored magnetic energy into heat and bulk flow.
Formalized by Furth, Killeen, and Rosenbluth in 1963, the tearing mode is a resistive instability driven by the free energy of the equilibrium current, gated by a purely geometric stability index Δ′, and growing on a hybrid timescale intermediate between the fast Alfvén time and the glacially slow resistive-diffusion time.
- RegimeResistive MHD; low-frequency, long-wavelength (ka ≲ 1)
- Instability criterionΔ′ > 0 (jump in ψ′/ψ across the resistive layer)
- DiscoveredFurth, Killeen & Rosenbluth, 1963 (Phys. Fluids)
- Growth-rate scalingγ ∝ S^-3/5 (constant-ψ) ; layer width δ ∝ S^-2/5 a
- Realized inTokamaks (NTMs, sawteeth), solar flares, magnetotail, MRX/TREX lab experiments
- Matters forFusion confinement/disruptions, solar flares, reconnection topology change
Interactive visualization
Press play, or step through manually. The visualization is yours to drive — try it before reading on.
Watch the 60-second explainer
A condensed visual walkthrough — narrated, captioned, under a minute.
What it is and why it matters
In ideal MHD, Alfvén's frozen-flux theorem forbids any change in magnetic topology: field lines are permanently tied to fluid elements. This is why a current sheet — two anti-parallel field regions separated by a thin layer of current — can sit in perfect force balance indefinitely. The tearing mode is the resistive instability that shatters this idealization. When even a tiny resistivity η is allowed, the plasma near the sheet's center can slip across field lines, and the sheet reorganizes into a periodic chain of nested magnetic islands (magnetic O-points) separated by X-points where reconnection occurs.
The consequence is enormous. Reconnection changes topology and releases stored magnetic energy: it powers solar flares (∼10²⁵ J), coronal mass ejections, and magnetospheric substorms, and it governs the sawtooth crashes and confinement-degrading islands that limit tokamaks. The tearing mode is the canonical, analytically tractable gateway to all of reconnection physics — the simplest problem where topology change is derived from first principles rather than assumed.
The mechanism, step by step
Consider an equilibrium field B_y(x) = B_0 tanh(x/a) reversing across a sheet of half-width a. Perturb it with a helical flux ψ ∝ exp(iky + γt). Away from the sheet the plasma is a near-perfect conductor, so the perturbation obeys the ideal, marginally-stable Newcomb equation ψ″ = (k² + B_y″/B_y)ψ — an outer region where inertia and resistivity are both negligible.
The key is the resonant surface where k·B = 0 (here x = 0), at which the ideal solution is singular: field lines there are exactly perpendicular to k, so bending them costs no energy and the frozen-flux constraint can be broken cheaply. Ideal MHD cannot connect the two outer solutions across this point. Only a thin resistive inner layer — where resistive diffusion, plasma inertia, and the Lorentz force all balance — smooths the singularity, allows field lines to break and rejoin, and forms the island. The instability's free energy comes from the gradient of the equilibrium current; reconnection relaxes it, driving the island wider.
The Δ′ criterion, growth rate, and scales
The heart of the theory is the tearing stability index, the jump in the logarithmic derivative of the flux across the layer:
Δ′ ≡ [ψ′/ψ]₀₋⁰⁺ = (1/ψ) (dψ/dx)|₀₊ − (1/ψ)(dψ/dx)|₀₋
Δ′ is set entirely by the outer ideal equation — pure equilibrium geometry, wavenumber, and boundary conditions. The mode is unstable when Δ′ > 0: this measures the free energy available from reconnecting flux, which must exceed the elastic cost of bending field lines. For the Harris sheet, Δ′ = 2(1/(ka) − ka)/a, so long wavelengths (ka < 1) are unstable.
Matching the inner resistive layer to this jump gives the Furth–Killeen–Rosenbluth dispersion relation. In the constant-ψ regime, γ = [Γ(¼)/(2π Γ(¾))]^{4/5} (Δ′)^{4/5} / (τ_H^{2/5} τ_R^{3/5}). With τ_H = a/v_A the Alfvén time and τ_R = a²/η the resistive time, and S = τ_R/τ_H the Lundquist number, this is γ ∝ Δ′^{4/5} S^{−3/5}, on a hybrid timescale ∝ S^{3/5} τ_H. The resistive layer is razor-thin: δ ∝ S^{−2/5} a. For a tokamak S ∼ 10⁸ and the corona S ∼ 10¹²–10¹⁴, so δ can be centimetres in a metres-wide device.
How it is realized, measured, and observed
In tokamaks, tearing modes appear as rotating helical islands at rational surfaces (safety factor q = m/n, e.g. the q = 2/1 or 3/2 surfaces). They are detected as coherent, rotating perturbations in Mirnov coils (edge magnetic pickup loops), as flat spots on electron-cyclotron-emission (ECE) temperature profiles (the island interior is isothermal because heat short-circuits along the reconnected field), and in soft-X-ray tomography. The mode number and rotation frequency read directly off the coil array.
Dedicated laboratory reconnection experiments — MRX (Princeton), TREX/TESSER (Wisconsin), and VTF (MIT) — drive current sheets and measure the reconnection electric field, island formation, and layer width in situ. In space, the Magnetospheric Multiscale (MMS) mission has directly sampled electron-scale reconnection layers in Earth's magnetotail and magnetopause since 2015, resolving the diffusion region. On the Sun, hard-X-ray flare loop-top sources and the cusp geometry of eruptive flares are the observational fingerprints of tearing-driven reconnection.
Where it operates, and distinctions from related effects
The tearing mode is a resistive instability: unlike the ideal kink or ballooning modes (which grow on the Alfvén time, γτ_A ∼ 1, and require Δ′-independent ideal free energy), it needs finite η and grows far more slowly in the classical FKR limit. It is one member of the resistive-MHD family that includes resistive interchange and rippling modes.
Its most consequential relatives are secondary and nonlinear developments. When a current sheet is driven to the ultra-thin Sweet–Parker configuration (aspect ratio ∼ S^{1/2}), it becomes tearing-unstable to the plasmoid instability above a critical Lundquist number S_c ≈ 10⁴, with a growth rate that increases with S (γτ_A ∼ S^{1/4}) — a super-Alfvénic runaway that shatters the sheet into a hierarchy of plasmoids and yields fast, S-independent reconnection (Loureiro 2007; Bhattacharjee 2009). The 'ideal tearing' insight (Pucci & Velli 2014) is that any sheet thinning to a/L ∼ S^{−1/3} tears on the Alfvén time regardless of S — setting a universal ceiling on how thin current sheets can get before disrupting. In tokamaks, the neoclassical tearing mode (NTM) is the dominant variant: even when Δ′ < 0 (classically stable), the loss of pressure-driven bootstrap current inside a seed island provides an additional drive that lets the island grow, degrading confinement.
Applications, significance, and open questions
The tearing mode is central to magnetic confinement fusion. NTMs cap the achievable plasma pressure (the β-limit) and can trigger disruptions that dump megajoules onto the wall; ITER's operating scenario depends on suppressing them with real-time electron-cyclotron current drive (ECCD) that replaces the missing bootstrap current inside the island (the modified Rutherford equation governs the balance). Sawtooth crashes, driven by the m/n = 1/1 resistive internal kink, redistribute the core on the tearing-layer timescale.
In astrophysics, tearing and the plasmoid cascade are the leading explanation for why reconnection in the corona (S ∼ 10¹²) proceeds ∼10³–10⁴ times faster than laminar Sweet–Parker theory allows — the flare 'fast reconnection' problem. Open questions remain sharp: how the collisional (resistive) picture connects to collisionless reconnection where electron inertia and the Hall term set the layer; the role of 3D turbulence and stochastic field lines; whether the plasmoid-mediated rate ∼0.01 v_A truly universalizes; and how to predict Δ′ and NTM onset well enough for burning-plasma control. The humble current sheet remains one of plasma physics' richest frontiers.
| Regime | Condition | Growth rate scaling | Layer / island signature |
|---|---|---|---|
| Constant-ψ (FKR) | Small Δ′, Δ′δ ≪ 1 | γ ∝ Δ′^4/5 S^-3/5 (τ_H^-2/5 τ_R^-3/5) | δ ∝ S^-2/5 a; slow, thin single island |
| Non-constant-ψ | Large Δ′, Δ′δ ≳ 1 | γ ∝ S^-1/3 (resistive-internal-kink-like) | δ ∝ S^-1/3 a; faster growth |
| Fastest linear mode | k a ~ S^-1/4 | γ τ_A ~ S^-1/2 | Peak over the unstable spectrum |
| Plasmoid instability | S > S_c ≈ 10⁴ (Sweet–Parker sheet) | γ τ_A ~ S^1/4 (super-Alfvénic!) | Many islands, N_plasmoid ~ S^3/8 |
| Ideal tearing | Sheet aspect ratio a/L ~ S^-1/3 | γ τ_A ~ O(1), independent of S | Onset of fast, S-independent reconnection |
Frequently asked questions
Why can't ideal MHD produce a tearing mode?
In ideal MHD, Alfvén's frozen-flux theorem ties field lines permanently to fluid elements, so magnetic topology can never change and no island can form. The tearing mode requires finite resistivity η in a thin inner layer around the resonant surface (k·B = 0), where field lines slip relative to the plasma and reconnect. That is why it is classed as a resistive instability — set η to zero and the growth rate vanishes as η → 0 in the classical limit.
What exactly is Δ′ and why does its sign matter?
Δ′ is the jump in the logarithmic derivative of the perturbed flux, [ψ′/ψ], across the resistive layer, computed from the ideal outer equation. It measures the free magnetic energy released by reconnecting flux versus the energy cost of bending field lines. Δ′ > 0 means reconnection is energetically favorable and the mode is unstable; Δ′ < 0 means it is classically stable. Crucially Δ′ is purely a property of the equilibrium and boundary conditions — it does not depend on resistivity.
How does the growth rate scale with the Lundquist number?
In the constant-ψ (small-Δ′) FKR regime, γ ∝ Δ′^(4/5) S^(−3/5), where S = a v_A/η, and the resistive layer width scales as δ ∝ S^(−2/5) a. For large Δ′ (non-constant-ψ) the scaling steepens to γ ∝ S^(−1/3). The fastest-growing mode over the spectrum satisfies γτ_A ∼ S^(−1/2) at ka ∼ S^(−1/4). All these are slow compared with ideal instabilities (γτ_A ∼ 1).
What is the difference between the classical tearing mode and the neoclassical tearing mode (NTM)?
The classical tearing mode is driven by the equilibrium current gradient through Δ′ and requires Δ′ > 0. The NTM adds a toroidal, kinetic drive: the pressure gradient flattens inside a seed island, which suppresses the local bootstrap current, and the missing current reinforces the island. NTMs can therefore grow even when the classical Δ′ is negative, provided a seed island (often from a sawtooth or ELM) exceeds a threshold width. They are the dominant performance-limiting instability in high-β tokamaks.
What is the plasmoid instability and how does it relate to fast reconnection?
When a current sheet is driven to the very thin Sweet–Parker configuration (aspect ratio ∼ S^(1/2)), it becomes violently tearing-unstable once S exceeds a critical value S_c ≈ 10⁴. This plasmoid instability has a growth rate that increases with S (γτ_A ∼ S^(1/4)), fragmenting the sheet into a self-similar hierarchy of plasmoids. The result is a reconnection rate ∼0.01 v_A that is nearly independent of S, resolving why solar-flare reconnection is far faster than laminar Sweet–Parker theory predicts.
Who discovered the tearing mode and when?
Harold Furth, John Killeen, and Marshall Rosenbluth published the foundational linear theory in 1963 (Physics of Fluids), deriving the Δ′ matching condition and the S^(−3/5) growth-rate scaling — hence 'FKR modes.' The nonlinear island-saturation theory came from Paul Rutherford in 1973 (the Rutherford equation). The neoclassical variant was developed in the 1980s–90s, and the plasmoid/ideal-tearing revival dates to Loureiro (2007), Bhattacharjee (2009), and Pucci & Velli (2014).