Landau contours and collisionless damping
Connect causal initial data to the resonant contour, determine the damping sign, and distinguish phase mixing from irreversible collisional heating.
On this page
The question for today
What you’ll understand
- Specify the causal analytic continuation for exp(ikx − iωt).
- Determine weak damping or growth from the imaginary dielectric response.
- Explain how electric-field decay can coexist with reversible Vlasov dynamics.
A pole at a wave's phase velocity is not a license to choose whichever contour gives a preferred answer. The causal initial-value problem fixes how the integral is defined before it is continued to damped frequencies. Landau damping follows from that prescription and the velocity distribution near resonance. Its apparent irreversibility is subtler than friction: the electric field can decay while the collisionless distribution develops finer structure and the ideal total energy remains conserved.
Causality fixes which side of the pole is retained
Use exp(ikx − iωt) with positive real k. A one-sided time transform of an initial perturbation is first defined sufficiently high in the upper complex-frequency half-plane for convergence. There the pole v = ω/k lies above the real velocity axis, so the ordinary real-axis integral is nonsingular. The causal response at other frequencies is obtained by analytic continuation from that initial domain, rather than by defining a fresh integral independently at each root.
As the pole moves downward, deform the velocity contour to keep passing below it. At a real phase velocity approached from above, the integral equals its principal value plus an imaginary pole term. The sign depends on the denominator and Fourier convention; reversing only one convention is an error. Smoothness or analyticity conditions on equilibrium and initial data determine how far continuation and pole-dominated asymptotics are justified. A damped Landau pole is not automatically a normal mode of the original real-velocity operator.
A small imaginary root is a controlled expansion
Write the complex frequency as a real part plus i times gamma, so positive gamma means growth. If the imaginary dielectric part and gamma are small, expand the dispersion equation around a simple real root of its real part. The damping rate is minus the imaginary dielectric value divided by the real frequency derivative. This expansion requires a nonzero derivative and a separation between the oscillation and growth or decay timescales.
For the positive-frequency Langmuir branch of a non-drifting Maxwellian, the real dielectric derivative is positive. The Maxwellian slope at positive phase velocity is negative, which gives a positive imaginary dielectric part under the stated convention and therefore negative gamma. A positive resonant slope can supply energy to this positive-energy branch instead. That slope rule is not universal for every wave or equilibrium: the wave-energy sign and other resonant species also enter the denominator and total response.
Phase mixing relocates information
Different velocities accumulate different spatial Fourier phases during streaming. Their contributions can cancel in the velocity integral even when none has individually decayed. The density moment and its electric field may therefore become small while the distribution retains oscillatory velocity structure. An exactly solvable free-streaming Gaussian perturbation demonstrates this cancellation, but it must be distinguished from self-consistent Landau dynamics: the latter also includes electric forces and collective response.
Ideal Vlasov evolution preserves fine-grained entropy and can support echoes when appropriately prepared perturbations rephase. Weak collisions act efficiently on the fine velocity scales produced by mixing and can then produce irreversible entropy. A finite-grid calculation can imitate this smoothing unintentionally. Exponential field decay over a fitted interval should thus be checked against velocity resolution, initial perturbation smoothness, collisions, and the nonlinear trapping time. Linear collisionless damping is a precise asymptotic statement with assumptions, not a claim that all disturbances vanish forever.
Work it through
Read a damping rate with its sign
At a simple Langmuir root, a causal calculation gives εi = 0.04, ∂ωεr = 2.0 × 10⁻⁶ s, and ωr = 10⁶ rad/s. Estimate gamma and the field-amplitude e-folding time.
The weak-root expansion gives . The positive imaginary dielectric value gives decay for this positive derivative.
The relative rate is , supporting the weak-damping approximation. With exp(−iωt), the amplitude envelope is exp(γt).
The amplitude e-folding time is . A squared field signal has half that e-folding time.
Gamma is −2.0 × 10⁴ s⁻¹, with a 50 μs field-amplitude lifetime and a 25 μs squared-amplitude lifetime.
The living notebook
Explore the model
Interactive lab · spectral
Resolve fine scales without confusing their dynamics
Reset and set BOTH Mode count N to 2 and Coefficient decay p to 0; predict two spatial harmonics. Then Reset and set BOTH N to 8 and p to 0; predict finer spatial structure. The instrument displays Σ sin(nx − t)/nᵖ. Its harmonics share one temporal frequency, so it illustrates resolution of fine structure only and cannot produce velocity-dependent phase mixing, Landau damping, or plasma echoes.
- Sum at x = π/2
- 0.867
- Smallest wavelength
- 2π / 5
- Highest-mode coefficient
- 0.2
A finite sum of sin(nx − t)/nᵖ. Playback changes t; increasing the cutoff includes finer spatial scales. This is not a proof about a continuum limit or renormalization. Parameter changes pause playback and recompute the model at the current elapsed time. Reset restores both parameters and returns time to zero. Playback starts paused. Adjust either parameter, then choose Play or Step.
Look forSeeing small-scale structure requires resolution, while the physical law that creates it must be specified independently.
Pause & reason
Check your understanding
Your turn
Try it yourself
In the force-free Gaussian example, let k = 2 m⁻¹ and vT = 3 m/s. Find the time at which the density Fourier amplitude is its initial value divided by e. Is that time a self-consistent Landau lifetime?
Take a moment to make a prediction or work it out on paper. Open a hint when you need one.
Give me a hint
Set (kvTt)²/2 = 1 and distinguish the model used to derive the Gaussian envelope.
Compare with the solution
The condition gives . The velocity-dependent factors themselves have unit magnitude; their integral decreases through cancellation. This is not a Landau lifetime because the stipulated evolution omits the self-consistent electric force. Its Gaussian time dependence differs from a pole-dominated exponential collective field response.
Keep the idea
Take it with you
- Causal initial data select the Landau contour and its resonant imaginary term.
- Damping signs follow the time convention and the dielectric frequency derivative together.
- Collisionless field decay can preserve fine information that collisions or numerical smoothing later erase.
Read further
Original sources and deeper explanations to follow at your own pace.
Optional progress, saved in this browser when available. Revisit any lesson at any time.
Lesson not marked complete.