research28 min studyLesson 915 in Physics

A cosmic-ray spectrum includes its propagation history

Relate diffusion length and escape to source spectra without treating arrival directions as straight source pointers.

On this page

What you’ll understand

  • Distinguish a diffusion Green function from a ballistic trajectory.
  • State geometric and energy-loss assumptions behind escape scalings.
  • Explain why source slope and transport slope are not separately fixed by one spectrum.

A charged particle arriving at Earth has generally interacted with magnetic structure along its route. Its direction, energy and arrival time can all differ from a straight-line source prediction. A diffusion equation summarizes many scatterings statistically, provided the distribution is sufficiently isotropic on the scales of interest. We will separate what such an equation predicts about a particle population from what a detector records for individual showers or tracks.

Diffusion is a coarse-grained population description

Let n(E,position,t) be particle number per volume per energy in the local medium frame. A scalar, spatially uniform diffusion coefficient D is an approximation to transport in magnetic turbulence; a coherent ordered field can require a diffusion tensor. The diffusion limit needs many scattering events and scales longer than a scattering mean free path and time. Its formal Green function has a tail at arbitrarily large distance immediately, so the approximation cannot describe a relativistic causal front or very early nearly ballistic transport.

For a point impulse injecting N particles in a fixed energy bin, with no boundaries or energy losses, the three-dimensional solution is a normalized Gaussian. Its mean squared radial displacement is 6Dt, whereas one Cartesian component has variance 2Dt. The root-mean-square distance is a population spread, not the path length traveled by a particle. In particular, dividing the spread by time does not produce the particle speed, which can remain very close to c throughout the random walk.

∂n∂t=D∇2n,n(r,t)=N(4πDt)3/2exp⁡ ⁣(−r24Dt),⟨r2⟩=6Dt\frac{\partial n}{\partial t}=D\nabla^2 n,\qquad n(r,t)=\frac{N}{(4\pi Dt)^{3/2}}\exp\!\left(-\frac{r^2}{4Dt}\right),\qquad \langle r^2\rangle=6Dt
This equation describes one energy bin in an infinite homogeneous three-dimensional medium, with t>0, D in m² s⁻¹ and n in m⁻³. For a differential spectrum, replace N by injected number per energy and give n the corresponding energy⁻¹ unit.↔ Scroll wide equations horizontally.

Boundary escape can steepen an injected spectrum

A finite confinement region loses particles through its boundaries. For one-dimensional diffusion starting at the center between absorbing surfaces at ±H, the mean first exit time is H²/(2D). This particular coefficient is a geometric first-passage result, not a universal Galactic residence-time formula. A steady population supplied at rate Q and lost on a characteristic residence time has total number approximately Q times that time. If diffusion rises as rigidity to a power δ, escape becomes faster at higher rigidity.

At fixed charge and ultrarelativistic energy, rigidity is proportional to energy. A source rate proportional to E⁻ᵖ then produces a simple loss-free reservoir spectrum proportional to E⁻⁽ᵖ⁺δ⁾. The conclusion concerns a specified steady escape model. A local density from a thin source disk and an extended halo has additional geometry-dependent factors; doubling the halo size does not universally quadruple a measured local flux. Winds, reacceleration, nearby discrete sources and radiative electron losses can change the slope or introduce breaks without changing the injection mechanism.

texit(0)=H22D,D(R)=D0(RR0)δ,Nsteady(E)≃Q(E)texit(E)∝E−(p+δ)t_{\rm exit}(0)=\frac{H^2}{2D},\qquad D(\mathcal R)=D_0\left(\frac{\mathcal R}{\mathcal R_0}\right)^\delta,\qquad N_{\rm steady}(E)\simeq Q(E)t_{\rm exit}(E)\propto E^{-(p+\delta)}
The exit coefficient applies to the stated one-dimensional absorbing slab. ℛ is magnetic rigidity, with reference ℛ₀ in the same units. The final power law assumes fixed particle charge, E≈pc, steady injection and negligible competing losses; Nsteady is total number per energy in the reservoir.↔ Scroll wide equations horizontally.

Losses and detector response complicate source inference

Spallation changes nuclear composition, while synchrotron and Compton cooling can reshape electron spectra. For an energy-loss rate b(E)=−dE/dt>0, a transport equation adds the term ∂[bn]/∂E; catastrophic removal requires a separate sink. Secondary-to-primary abundance ratios constrain traversed material rather than directly naming a unique source distance. At ultrahigh energies, proton photopion losses, nuclear photodisintegration and cosmic expansion also limit propagation. Arrival direction alone is especially ambiguous for charged nuclei because magnetic deflection depends on rigidity and the intervening field. Air-shower energy and composition are inferred through shower and detector-response models, not read directly from the source’s injected spectrum.

The heat lab isolates one diffusion mechanism. Reset and set both controls to mode one and diffusivity 0.5; Step twice to time one, predicting amplitude exp(−0.5)≈0.607. Reset and set mode two and diffusivity 0.5; Step twice, predicting exp(−2)≈0.135. Shorter spatial structure fades faster at the same diffusivity. This analytic periodic one-dimensional Fourier mode is not the three-dimensional point-source Green function, a random particle realization, an energy-loss calculation or a detector response model.

Calculate a three-dimensional diffusion spread

For a hypothetical loss-free energy bin, take D=10²⁴ m²/s and elapsed medium-frame time t=10¹² s. Find the radial root-mean-square displacement and express it in parsecs using 1 pc=3.086×10¹⁶ m. State what this length represents.

  1. The mean squared radial displacement is 6Dt=6×10³⁶ m². The factor six accounts for three independent Cartesian variances.

  2. Taking the square root gives √6×10¹⁸≈2.45×10¹⁸ m. This length grows as the square root of time.

  3. Divide by 3.086×10¹⁶ m/pc to obtain about 79.4 pc. The result is the population spread around the injection point, not the much larger cumulative trajectory length.

The result

The model spread is about 2.45×10¹⁸ m, or 79.4 pc, under the homogeneous diffusion assumptions.

Explore the model

Interactive lab

Interactive lab · heat

Check the diffusion rate of spatial structure

Reset and set BOTH controls n=1,κ=0.5; Step twice to t=1: predict amplitude e⁻⁰·⁵≈0.607. Reset and set n=2,κ=0.5; Step twice: predict e⁻²≈0.135. This is a periodic one-dimensional exact Fourier mode, with no cosmic-ray trajectories, energy bins or source boundaries.

Check the diffusion rate of spatial structureThe exact periodic heat-equation mode is u(x,t) = exp(−κn²t) sin(nx), on [0,2π]. Playback advances physical time. The initial sine is shown separately, and the unnormalized L² norm is √π exp(−κn²t). This analytic instrument does not run a finite-difference solver or test discretization stability. Current readings: Current amplitude: 1; Spatial L² norm: 1.772; Decay rate · κn²: 3.6 /s. All parameters are adjustable with the labeled controls below.Periodic heat equation mode 3 at time 0: the amplitude is 1.0π/2π3π/22π−1−0.500.51x · periodic intervalTemperature perturbation uExact mode · n = 3
Current temperature modeInitial sineCurrent amplitude
Current amplitude
1
Spatial L² norm
1.772
Decay rate · κn²
3.6 /s
t = 0 s

The exact periodic heat-equation mode is u(x,t) = exp(−κn²t) sin(nx), on [0,2π]. Playback advances physical time. The initial sine is shown separately, and the unnormalized L² norm is √π exp(−κn²t). This analytic instrument does not run a finite-difference solver or test discretization stability. 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 forDiffusion suppresses larger wave numbers faster, while a cosmic-ray propagation model needs additional geometry and loss physics.

Check your understanding

An observed loss-free steady reservoir spectrum has index 2.6. Which inference is justified in the simple escape model?

Try it yourself

In a hypothetical steady absorbing slab with injection at its center, let injection index p=2.2 and diffusion index δ=0.4. Find the total-population energy index. Then double H while keeping D, the total injection rate and all loss assumptions fixed; find the change in mean residence time and total stored population.

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 p+δ for the index and H²/(2D) for the stated first-exit time. The question concerns integrated reservoir population, not local detector flux.

Compare with the solution

The energy index is 2.2+0.4=2.6. Doubling H multiplies the first-exit time by four and therefore multiplies the integrated steady population by four at each energy. The index stays 2.6 if the same rigidity dependence applies. Inferring a local flux would additionally require the spatial source distribution and boundary solution.

Take it with you

  • Diffusion predicts a population distribution rather than an individual path.
  • Escape coefficients and local densities depend on source and boundary geometry.
  • An observed spectral slope can combine injection, transport and losses.

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.