Fermi Acceleration
Intuition
Bounce a ping-pong ball between two paddles moving slowly together: every bounce off an approaching paddle adds speed. Replace paddles with magnetic mirrors and the ball with a trapped charged particle, and you have Fermi's 1949 mechanism for accelerating cosmic rays to fantastic energies — magnetic clouds in the galaxy acting as drifting mirrors.
The setup
The course's model: a mirror field whose length slowly shrinks,
A time-varying \(\mathbf{B}\) induces an electric field (Faraday), computed from the vector potential:
This azimuthal E field is what actually does the work — magnetic fields never do.
The physics
- Each reflection off an approaching mirror boosts \(v_\parallel\) (head-on collision with a moving wall: \(v \to v + 2u\) in the wall frame).
- The parallel invariant \(J = \oint v_\parallel\, dz \approx v_\parallel L\) is adiabatically conserved: as \(L\) shrinks, \(v_\parallel\) must grow — \(v_\parallel \propto 1/L\).
- Meanwhile \(\mu = mv_\perp^2/2B\) holds \(v_\perp\) tied to \(B\). The pitch angle therefore decreases as \(v_\parallel\) grows...
- ...until the particle enters the loss cone (\(\theta < \theta_\text{trap} = \sin^{-1}(\lambda^{-1})\) in the model's notation) and escapes — carrying its stolen energy. Acceleration with a built-in exit door: exactly what a cosmic-ray source needs.
Numerical experiment
Integrate the full orbit with the leapfrog solver, fields from the analytic derivatives of \(\psi\) (chain rule through \(L(t)\)):
Watch \(v_\parallel\) ratchet up bounce after bounce, then the escape. Track \(\mu\) to confirm the perpendicular invariant holds while \(J\) does the accelerating.
Where it matters
Cosmic-ray acceleration (Fermi's original 2nd-order mechanism, and the 1st-order version at supernova shocks) · magnetic-pumping heating schemes · particle energization in shrinking magnetospheric flux tubes.
Common mistakes
- Crediting the magnetic field with the energy. The energy comes from whatever moves the mirrors, delivered via the induced E field.
- Expecting unlimited acceleration. The loss cone always wins eventually; the mechanism accelerates and releases.
- Simulating with a symplectic-but-static-field mindset: time-dependent fields mean energy is not conserved — that's the whole point; don't "fix" it.
Related concepts
- Magnetic mirror — the static prerequisite
- Guiding center & drifts — adiabatic framework
- Leapfrog method — integration engine
Knowledge graph position
Prerequisites: Magnetic mirror. Leads to: astrophysical particle acceleration, wave heating.
Quiz
Q1 (computational). Mirrors shrink from \(L_0\) to \(L_0/2\). By what factor does a trapped particle's parallel energy grow (adiabatic \(J\))?
Answer
\(v_\parallel \propto 1/L\) doubles ⇒ parallel energy ×4.
Q2 (conceptual). Why does Fermi acceleration preferentially increase energy, when a particle can also bounce off a receding mirror and lose energy?
Answer
Head-on (approaching) encounters are more frequent than overtaking ones — approach speed is higher, so collisions are biased toward gains (rate ∝ relative velocity). Statistically the gains win — 2nd-order Fermi. With converging mirrors (or a shock), every bounce gains — 1st-order Fermi.
Q3 (MCQ). The work on the accelerated particle is done by:
- (a) the magnetic field directly (b) the induced electric field
- (c) collisions (d) gravity
Answer
(b). \(\partial\mathbf{B}/\partial t \neq 0\) induces \(\mathbf{E}\) (Faraday); only \(\mathbf{E}\) does work on charges.