Evolve angular radiation moments with gravity and collisions
Separate Fourier wave number, angular multipoles, tight coupling, and hierarchy closure.
On this page
The question for today
What you’ll understand
- Derive neighboring angular-moment coupling for a collisionless Fourier mode.
- State the expansion and scattering hierarchy required for tight coupling.
- Distinguish internal radiation multipoles from an observed sky power spectrum.
Radiation has a direction-dependent distribution, so density and velocity alone do not close its evolution after particles stream freely. Expanding the angular dependence in Legendre moments converts a linear Boltzmann equation into a coupled hierarchy. Gravity, scattering, polarization, and the matter background then supply additional terms and constraints. A cosmological prediction solves that system with consistent initial conditions and a controlled numerical truncation. The word hierarchy does not imply that every higher moment is negligible, and its angular index is not simply a spatial Fourier cutoff.
Expand directions around one spatial wave vector
For a scalar Fourier mode, let μ be the cosine between photon direction and k, and Θ its fractional temperature disturbance. Expand Θ in Legendre polynomials with coefficients (−i)ℓ(2ℓ+1). That phase convention determines the signs in the moment equations. In a collisionless test problem without metric driving, Θ′+ikμΘ=0. The Legendre recurrence then couples each multipole only to its two neighbors, transferring structure into progressively finer angular variation.
The monopole describes the angular average, the dipole relates to bulk motion, and the quadrupole contributes to anisotropic stress. Factors depend on whether one expands temperature, intensity, or a momentum-dependent distribution. In particular, a photon intensity perturbation is four times the temperature perturbation at linear order. Massive neutrinos require momentum dependence because their speed changes with momentum and expansion; replacing them by one massless temperature hierarchy would discard essential information.
Recover an acoustic fluid only in tight coupling
Thomson scattering couples photons to free electrons and hence baryons. Define the positive conformal scattering rate Γ=a neσT, rather than silently using a signed optical-depth derivative. Tight coupling requires Γ much larger than both the relevant wave number and conformal expansion rate, together with slowly varying coefficients for the expansion used. The nearly common photon–baryon velocity then supports acoustic oscillations with a sound speed reduced by baryon inertia.
Finite mean free path generates slip, higher moments, polarization, and diffusion damping. A controlled calculation expands in the small scattering time and switches treatment when that expansion loses accuracy. Recombination changes the free-electron density and therefore the collision history; it is not simply an abrupt deletion of every scattering term at one fixed time. Free-streaming neutrino anisotropic stress also feeds metric potentials, so setting Φ=Ψ throughout radiation evolution would generally remove a physical source.
Project to the sky and test numerical closure
The full linear system couples radiation moments, species densities and velocities, and Einstein constraints. Its solution at each k supplies transfer functions projected onto observed sky multipoles. For a single adiabatic scalar initial mode, the sky power integrates primordial power times the squared radiation transfer. Other initial modes and correlations add cross terms. An internal hierarchy multipole at one time is not automatically the observed sky coefficient; propagation and projection connect them. Artificially setting all higher moments to zero can reflect errors back toward lower multipoles as angular structure reaches the cutoff. Increase the maximum multipole and test the closure and time integration; apparent convergence of one observable at one k does not certify every scale.
The spectral lab offers a separate cutoff comparison. Reset and set mode count 2,decay exponent 1; at time 0 predict sum at x=π/2 equal to 1 and smallest wavelengthπ. Reset and set count 4,exponent 1; predict sum 2/3 and smallest wavelengthπ/2. Its finite traveling sine sum is prescribed, not a Legendre hierarchy. It contains no gravitational sources, collisions, polarization, or line-of-sight projection; changing its cutoff cannot validate a cosmological hierarchy closure.
Work it through
Compute three instantaneous free-streaming derivatives
In the undriven collisionless hierarchy take k=0.3 Mpc⁻¹ and moments Θ0=0.010,Θ1=0.002,Θ2=0.001,Θ3=0. Compute Θ0′,Θ1′ and Θ2′ at that instant.
The monopole equation gives Θ0′=−kΘ1=−0.3×0.002=−0.0006 Mpc⁻¹. Its value can change despite no collisions because angular transport couples it to the dipole.
The dipole equation gives Θ1′=(k/3)(Θ0−2Θ2)=0.1×(0.010−0.002)=0.0008 Mpc⁻¹.
The quadrupole equation gives Θ2′=(k/5)(2Θ1−3Θ3)=0.06×0.004=0.00024 Mpc⁻¹. These are local derivatives, not a complete finite-time solution with higher moments artificially held fixed.
The derivatives are −6×10⁻⁴,8×10⁻⁴ and 2.4×10⁻⁴ per Mpc in the stated conformal-time convention.
The living notebook
Explore the model
Interactive lab · spectral
Compare two spatial cutoffs
Reset and set N=2,p=1; at t=0 predict sum 1 at x=π/2 and smallest wavelengthπ. Reset and set N=4,p=1; predict sum 2/3 and smallest wavelengthπ/2.
- 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 forThis traveling sine series has a spatial cutoff, not an angular Boltzmann closure. It does not evolve radiation, polarization, or CMB transfer.
Pause & reason
Check your understanding
Your turn
Try it yourself
Assume Rb=1/2 is constant during a short tightly coupled interval Δη=3 Mpc, with c=1. Find cs and the sound-horizon increment. Why is the interval assumption needed?
Take a moment to make a prediction or work it out on paper. Open a hint when you need one.
Give me a hint
Use cs²=1/[3(1+Rb)] and integrate cs over conformal time.
Compare with the solution
The squared speed is 2/9, so cs=√2/3≈0.4714. The sound-horizon increment is csΔη=√2 Mpc≈1.414 Mpc. Treating Rb as constant makes this integral elementary; in a physical expanding background Rb and the coupling rate evolve, so a full history is required and tight coupling must remain valid throughout.
Keep the idea
Take it with you
- Direction-dependent radiation needs an angular hierarchy after free streaming.
- Tight coupling is a controlled rate expansion, not a permanent fluid identity.
- Observed sky power requires physical evolution, projection, and numerical convergence checks.
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.