Spotlight #36 – Plasma Physics & Fusion: Tokamak Confinement, Lawson Criterion, MHD Stability and ICF

Ninety-nine percent of all visible matter in the universe exists in the plasma state. The sun fuses 620 million tonnes of hydrogen per second by confining its core plasma with gravity at 15 million K. On Earth, both magnetic confinement fusion (MCF) in tokamaks and inertial confinement fusion (ICF) aim to replicate this process without gravitational assistance — one of the hardest engineering challenges humanity has attempted. Understanding why requires a tour of plasma physics from first principles.

1. Plasma: The Fourth State of Matter

A plasma is a quasi-neutral gas of charged particles (electrons and ions) exhibiting collective behaviour. Three criteria distinguish a plasma from an ionised gas:

Debye length (shielding distance):
  λ_D = √(ε_0 k_B T_e / (n_e e²))
  For T_e = 10 keV, n_e = 10²&sup0; m−³: λ_D ≈ 74 μm

Plasma criteria:
  1. λ_D ≪ L  (system size much larger than shielding length)
  2. N_D = n_e · (4π/3) λ_D³ ≫ 1  (many electrons in Debye sphere)
  3. ω_pe · τ ≫ 1  (plasma oscillation period shorter than collision time)

Plasma (electron) frequency:
  ω_pe = √(n_e e² / (ε_0 m_e))
  For n_e = 10²&sup0; m−³: ω_pe ≈ 5.6 × 10¹¹ rad/s  (microwave range)

Cyclotron frequencies:
  Ω_ce = eB / m_e   (electron)
  Ω_ci = eB / m_i   (ion)   Ω_ci ≈ Ω_ce / 1836  (proton)
          

The Debye length is the screening length over which electric fields are shielded by the redistribution of free charges. Beyond λ_D, a plasma looks electrically neutral. This collective shielding is what makes plasma qualitatively different from a weakly ionised gas and gives rise to the rich wave phenomena discussed in section 5.

2. Tokamak Magnetic Confinement

The tokamak (from Russian: тороїдальна камера з магнітними котушками) confines plasma in a torus using a combination of a strong toroidal magnetic field B_φ from external coils and a weaker poloidal field B_θ generated by a large plasma current. The resultant helical field lines wind around the torus without intersecting the wall.

Tokamak geometry:
  R = major radius (axis to torus centre)
  a = minor radius (cross-section radius)
  Aspect ratio: A = R/a   (ITER: R = 6.2 m, a = 2.0 m, A = 3.1)

Safety factor q (field-line winding):
  q = r B_φ / (R B_θ) ≈ rB_T / (RB_P)
  Kruskal-Shafranov condition for stability: q ≥ 1 everywhere
  q_95 ≈ 3 for ITER design point

Toroidal current: I_P ≈ 15 MA  (ITER nominal)

Neoclassical transport (from banana orbits):
  D_⊥_neo ~ (q/ε)^(3/2) ρ_i² v_thi / λ_mfp²
  Bohm-like anomalous diffusion: D_Bohm = k_BT / (16eB)
  Anomalous dominates; actual confinement time τ_E ≈ 1–3 s (ITER design)

L-H transition (H-mode):
  At input power P_th ~ B_T n_e^0.7 R^2: edge transport barrier forms
  Pedestal: steep density/temperature gradients
  ELM (edge-localised mode): periodic pedestal crashes, heat loads
          

Particle drifts in the curved, inhomogeneous tokamak field — grad-B drift, curvature drift, and E×B drift — would cause plasma to drift outward and hit the wall in microseconds if the field were purely toroidal. The twist provided by the poloidal field averages these drifts over each orbit, maintaining confinement. Particles following helical field lines trace “banana orbits” in the poloidal cross-section; their collisional diffusion gives the neoclassical transport estimate.

3. The Lawson Criterion

Fusion of deuterium and tritium (D-T) releases 17.6 MeV per reaction, primarily as a 14.1 MeV neutron and a 3.5 MeV alpha particle. The alpha remains confined and heats the plasma; the neutron escapes and its energy is extracted as heat. For a fusion reactor to produce more energy than it consumes (ignition), the plasma heating by alpha particles must exceed all energy loss channels.

D-T fusion reaction:
  D + T → &sup4;He (3.5 MeV) + n (14.1 MeV)
  Peak reactivity at T ≈ 70 keV; ⟨σv⟩_max ≈ 3.7 × 10−²² m³/s  (at ~65 keV)

Power balance (ignition condition):
  P_α = n_D n_T ⟨σv⟩ E_α / 4   (alpha heating)
  P_loss = 3 n k_B T / τ_E   (Bremsstrahlung + conduction)

Lawson criterion (ignition point for D-T):
  n τ_E ≥ 1.5 × 10²&sup0; m−³ s   (at T ≈ 25 keV)

Fusion triple product (Lawson + temperature):
  n T τ_E ≥ 3 × 10²¹ m−³ keV s

ITER design targets:
  n ≈ 10²&sup0; m−³,  T ≈ 15 keV,  τ_E ≈ 3.7 s
  Q = P_fusion / P_input = 10  (10× energy multiplication)

Current record (JET, 2022): Q ≈ 0.33
NIF ignition (2022): Q_target = 1.5  (→ fusion energy > laser energy to capsule)
          

4. MHD Stability

Magnetohydrodynamics (MHD) treats plasma as a conducting fluid. The ideal MHD equations combine Maxwell’s equations with the fluid momentum equation, subject to the frozen-in-flux condition. MHD instabilities can expel hot plasma from the confinement region on timescales of microseconds if unchecked.

Ideal MHD momentum equation:
  ρ dv/dt = J × B − ∇p   (Lorentz force + pressure gradient)

Frozen-in-flux (Alfvén 1942):
  ∂B/∂t = ∇ × (v × B)
  Field lines move with the plasma; resistivity breaks this (tearing modes)

Kink instability (m=1):
  Criterion (Kruskal-Shafranov): q < 1 unstable
  Long-wavelength current-driven; bends entire plasma column

Sausage instability (m=0):
  Constriction of plasma column;  stable if B_φ² > B_θ²/2

Tearing mode (resistive MHD):
  Reconnects field lines at rational surfaces q = m/n
  Growth rate: γ ~ (k_⊥ v_A)^(3/5) (η/μ_0)^(2/5) r_s^(−2/5)
  Islands of disconnected flux; can lock to wall → disruption

ELMs (Edge-Localised Modes):
  Peeling-ballooning instability in H-mode pedestal
  Type-I ELMs: ΔW_ELM ≈ 1–20 MJ thrown to divertor
  ITER mitigation: pellet injection, resonant magnetic perturbations
          

5. Plasma Waves

Because plasma is both electromagnetic and hydrodynamic, it supports a rich zoo of wave modes that do not exist in ordinary gases. Many of these are exploited for heating: ion cyclotron resonance heating (ICRH), electron cyclotron resonance heating (ECRH), and lower-hybrid current drive are all used in tokamaks.

Alfvén wave (MHD wave propagating along B):
  v_A = B / √(μ_0 ρ)
  For B = 5 T, n = 10²&sup0; m−³ (D): v_A ≈ 1.1 × 10&sup7; m/s ≈ 4% c

Shear Alfvén: ω = k_∥ v_A  (transverse oscillation of field lines)
Compressional Alfvén: ω² = k² v_A²  (fast magnetosonic, isotropic)

Whistler wave (right-hand circularly polarised, f < f_ce):
  Propagates along B; group velocity v_g ∝ √f
  High frequencies arrive earlier → dispersed whistle in ionosphere

Upper/Lower Hybrid resonances:
  ω_UH² = ω_pe² + Ω_ce²   (ECRH heating layer)
  ω_LH² = Ω_ci Ω_ce · (ω_pi² + Ω_ci²) / (ω_pi² + ω_pe² + Ω_ci²)

ICRH heating:
  ω = Ω_ci = eB/m_i   (ion cyclotron resonance)
  Minority species heating: H minority in D plasma absorbs power
  Power coupled: P ≈ 10–20 MW per antenna system in ITER
          

6. Inertial Confinement Fusion

Inertial confinement fusion (ICF) uses intense laser pulses (or X-rays from a laser “hohlraum”) to ablate the outer shell of a millimetre-scale fuel capsule. The rocket reaction compresses the D-T fuel to densities thousands of times that of solid matter, simultaneously shock-heating the central “hot spot” to fusion temperatures. The plasma is confined inertially — by its own mass — for the nanoseconds before it disassembles.

ICF drive requirements (indirect drive):
  Hohlraum X-ray radiation temperature: T_r ≈ 300 eV
  Laser energy: E_L ≈ 2.05 MJ (NIF, 192 beams, 351 nm)
  Capsule absorbed: ~150 kJ

Compression:
  Convergence ratio CR = R_initial / R_final ≈ 35
  Final density: ρ ≈ 1000 g/cm³  (≈ 400× solid DT)
  Hot-spot temperature: T_hs ≈ 10 keV (100 million K)

Ignition condition (hot-spot):
  ρR_hs ≥ 0.3 g/cm²   (confinement parameter)

Rayleigh-Taylor instability (main limiting factor):
  Grows at ablation front: γ = √(A_t k g) where A_t = Atwood number
  Requires surface roughness < 50 nm RMS on capsule

NIF ignition milestone (December 2022):
  Laser energy in: 2.05 MJ
  Fusion energy out: 3.15 MJ   (Q = 1.54 → scientific ignition)
  Confinement time: τ ≈ 100 ps
  First laboratory thermonuclear ignition in history
          

Try These Simulations