Thomson and Compton Scattering

Electron scattering is a working tool across modern astrophysics. The Planck satellite mapped the polarization of the cosmic microwave background, which exists only because Thomson scattering at last scattering ($z\approx1090$) imprints linear polarization on a quadrupole radiation field; the E-mode and B-mode patterns now probe reionization and primordial gravitational waves. The Sunyaev–Zel'dovich effect has matured into a survey engine: the South Pole Telescope, ACT, and Planck have catalogued thousands of clusters by their $y$-decrement, yielding redshift-independent masses that constrain structure growth and neutrino masses. In the X-ray band, NuSTAR dissects the Comptonized power-law continua of accreting black holes — reading corona temperatures and optical depths from the cutoff near a few hundred keV — while IXPE measures the polarization those coronae imprint. And the electron-scattering opacity remains the backbone of stellar and disk modeling, fixing the Eddington limit that caps the luminosity of massive stars, quasars, and tidal-disruption flares.

Start with the softest possible photon and a single free electron. The wave's field $E_0\cos\omega t$ shakes the electron along the field direction; being free, the electron has no restoring force and no resonance, so its acceleration amplitude $a_0=eE_0/m_e$ carries no memory of the drive frequency. An accelerating charge radiates by the Larmor formula, and the reemitted power divided by the incident flux is a pure constant — $\sigma_T$ — whose frequency independence is not an approximation but an exact statement about a charge with no natural frequency. The radiated power is not isotropic: the dipole pattern averages, for unpolarized light, to the $(1+\cos^2\theta)$ shape below, peaked forward and backward and halved at right angles. That angular asymmetry, acting on a quadrupole anisotropy in the radiation field, is exactly what polarizes the cosmic microwave background.

Now let the photon hit harder. An X-ray or gamma ray carries momentum $h\nu/c$ comparable to $m_e c$; the electron recoils, and momentum conservation forces the scattered photon to lower energy. The harder the hit, the more energy is lost and — the quantum surprise the classical picture cannot give — the smaller the effective cross section becomes. The second figure is the punchline of the subject: the total cross section versus the dimensionless photon energy $\varepsilon = h\nu/m_ec^2$. It sits flat on the Thomson plateau while $\varepsilon\ll1$, then Klein–Nishina bends it down as the energy nears $511\ \mathrm{keV}$, falling as $\sim1/\varepsilon$ so the most energetic gamma rays see a nearly transparent electron.

Finally, energy can flow either way. If the photon is more energetic, it loses energy — ordinary Compton. If the electron is hotter, the photon gains energy on average — inverse Compton. Neither changes the cross section; "inverse" names only the direction of net energy transfer. Repeated scatterings in a hot medium push a photon gas toward equilibrium with the electrons, reshaping spectra in a process called Comptonization.

The cornerstone is the Thomson cross section itself, built from the classical electron radius $r_e$ — an area of order the square of a nuclear size, which is why electron scattering is so feeble that only very ionized or very extended plasmas make it matter.

where: $\sigma_T$ — total Thomson scattering cross section per free electron; $r_e = e^2/(4\pi\varepsilon_0 m_e c^2) = 2.818\times10^{-15}\ \mathrm{m}$ — classical electron radius. Reading it: this is the geometric target area a free electron presents to soft light, fixed once and for all by the electron's charge and mass. The scattering is not isotropic. Averaging the dipole pattern over unpolarized light gives the angular distribution, whose integral over solid angle reproduces equation (1); its forward–backward symmetry and $\theta=90°$ minimum make Thomson scattering a source of linear polarization.

where: $\theta$ — angle between the incident and scattered photon directions; $\mathrm{d}\Omega$ — solid-angle element of the scattered beam. Reading it: twice as much light scatters straight forward or straight back as sideways, and $\int(\mathrm{d}\sigma/\mathrm{d}\Omega)\,\mathrm{d}\Omega=\sigma_T$. When the photon energy is no longer negligible the electron recoils. Conserving relativistic energy and momentum for a photon striking a stationary electron gives the Compton shift — a wavelength change that depends only on the scattering angle, not on the incident wavelength.

where: $\lambda,\ \lambda'$ — incident and scattered wavelengths; $\lambda_C$ — Compton wavelength of the electron; $\theta$ — scattering angle. Reading it: the wavelength grows by at most $2\lambda_C$ (at back-scatter), an absolute shift that is negligible for optical light but order-unity for hard X-rays and gamma rays. The same physics written in energies is often more convenient, since it displays the dimensionless energy $\varepsilon$ that controls the whole crossover from Thomson to Klein–Nishina.

where: $E,\ E'$ — incident and scattered photon energies; $\varepsilon$ — photon energy in units of $m_ec^2=511\ \mathrm{keV}$. Reading it: for $\varepsilon\ll1$ the ratio is $\approx1$ (elastic Thomson), while for $\varepsilon\gtrsim1$ the back-scattered photon loses most of its energy. The exact quantum cross section per solid angle is the Klein–Nishina formula, derived from the Dirac equation. It reduces to the dipole result (2) as $E'\to E$ and increasingly favors forward scattering as the energy rises.

where: $E'/E$ — the Compton energy ratio from equation (4); $\theta$ — scattering angle. Reading it: the extra factor $(E'/E)^2$ suppresses the large-angle scattering that loses the most energy, tilting the pattern forward. Integrating (5) over solid angle gives the total Klein–Nishina cross section, the curve plotted in the second figure. It is the exact generalization of $\sigma_T$ to all energies.

where: $\sigma_{KN}\to\sigma_T$ as $\varepsilon\to0$, and $\sigma_{KN}\approx(3/8)\sigma_T\,\varepsilon^{-1}(\ln2\varepsilon+\tfrac12)$ as $\varepsilon\to\infty$. Reading it: the cross section is flat at low energy and falls as $\sim1/\varepsilon$ at high energy, softening the scattering. For astrophysics the per-electron cross section is most useful as a mass opacity, got by counting free electrons per unit mass of ionized gas — the grey opacity that dominates hot stellar interiors.

where: $\kappa_{es}$ — electron-scattering mass opacity; $n_e$ — free-electron number density; $\rho$ — mass density; $X$ — hydrogen mass fraction (full ionization assumed). Reading it: for solar composition $X=0.7$, $\kappa_{es}\approx0.034\ \mathrm{m^2\,kg^{-1}}=0.34\ \mathrm{cm^2\,g^{-1}}$, and this floor on the opacity sets the Eddington limit. Finally, in a hot medium the net energy exchanged per scattering has a recoil term (photon loses) and a Doppler term (photon gains from the moving electron); averaged over a thermal electron population it gives the Comptonization kernel.

where: $T_e$ — electron temperature; $h\nu$ — photon energy; $k$ — Boltzmann constant. Reading it: photons gain energy when $h\nu<4kT_e$ (inverse Compton, the SZ regime) and lose it when $h\nu>4kT_e$; the two balance at the Compton equilibrium $h\nu=4kT_e$.

The whole subject is organized by the single dimensionless energy $\varepsilon = h\nu/m_ec^2$, which measures the photon energy against the electron rest energy and controls the crossover from classical to quantum scattering.

1 · Eddington luminosity of a hot star (fully worked)

A one-solar-mass star has a fully ionized envelope of solar composition $X=0.7$, where electron scattering dominates the opacity. Find the Eddington luminosity — where outward radiation pressure on the electrons just balances gravity — and compare it to the Sun's output.

Step 1. Opacity from equation (7): $\kappa_{es} = 0.02(1+X) = 0.02\times1.7 = \mathbf{0.034\ m^2\,kg^{-1}}$. Bound processes only add to this, so it is a firm lower bound on the total opacity.

Step 2. Balancing radiative force per unit mass $\kappa_{es}F/c$ against gravity $GM/r^2$, the $r^{-2}$ cancels, giving $L_{Edd} = 4\pi G M c/\kappa_{es}$. The numerator is $4\pi G M c = 4\pi(6.674\times10^{-11})(1.989\times10^{30})(2.998\times10^8) = 5.00\times10^{29}$ (SI).

Step 3. Divide: $L_{Edd} = (5.00\times10^{29})/0.034 = \mathbf{1.47\times10^{31}\ W}$, or with $L_\odot = 3.828\times10^{26}\ \mathrm{W}$, $L_{Edd} \approx \mathbf{3.8\times10^{4}\ L_\odot}$. The Sun sits $40{,}000$ times below its Eddington limit — hence stable — but a $100\,M_\odot$ star, whose luminosity outpaces its mass, sits close to $L_{Edd}$ and drives powerful radiation-driven winds. The Eddington luminosity is $\approx1.5\times10^{31}\ \mathrm{W}\approx3.8\times10^{4}\,L_\odot$ per solar mass; the constancy of $\sigma_T$ makes this limit so clean.

2 · The Compton $y$-parameter of a galaxy cluster (fully worked)

CMB photons pass through the hot intracluster gas of a massive cluster: electron temperature $T_e = 10^8\ \mathrm{K}$, electron density $n_e = 10^{3}\ \mathrm{m^{-3}}$ (i.e. $10^{-3}\ \mathrm{cm^{-3}}$), path length $L = 1\ \mathrm{Mpc} = 3.09\times10^{22}\ \mathrm{m}$. Estimate the Thomson optical depth, the Compton $y$-parameter, and the Sunyaev–Zel'dovich temperature decrement in the Rayleigh–Jeans part of the CMB.

Step 1. Thomson optical depth $\tau = n_e\sigma_T L = (10^{3})(6.65\times10^{-29})(3.09\times10^{22}) = \mathbf{2.1\times10^{-3}}$ — about one CMB photon in five hundred scatters in the cluster.

Step 2. The thermal boost per scattering is $kT_e/m_ec^2$. With $kT_e = (1.38\times10^{-23})(10^8)/(1.60\times10^{-16}) = 8.6\ \mathrm{keV}$, this is $8.6/511 = \mathbf{0.0169}$.

Step 3. The Compton parameter is the product $y = (kT_e/m_ec^2)\,\tau = 0.0169\times2.1\times10^{-3} = \mathbf{3.5\times10^{-5}}$ — the integrated dimensionless electron pressure that SZ surveys map.

Step 4. In the Rayleigh–Jeans limit $\Delta T/T = -2y$, so $\Delta T = -2(3.5\times10^{-5})(2.725\ \mathrm{K}) = \mathbf{-190\ \mu K}$: a decrement below $218\ \mathrm{GHz}$, turning to an increment above it. A rich cluster gives $y\sim10^{-4}$ and a decrement of a few hundred $\mu\mathrm{K}$, independent of redshift — what makes the SZ effect a redshift-blind cluster finder.

3 · When does recoil matter?

Compare the back-scatter shift for a $500\ \mathrm{nm}$ optical photon and a $100\ \mathrm{keV}$ hard X-ray.

Step 1. The maximum shift $\Delta\lambda = 2\lambda_C = 4.85\times10^{-12}\ \mathrm{m}$ is the same absolute amount for both. Optically, $\Delta\lambda/\lambda = 4.85\times10^{-12}/5.0\times10^{-7} = \mathbf{1\times10^{-5}}$ — negligible, so optical scattering off electrons is pure Thomson.

Step 2. For the X-ray, $\varepsilon = 100/511 = 0.196$, and the energy form (4) gives $E' = 100/(1+2\times0.196) = \mathbf{72\ keV}$, a $28\%$ loss in one back-scatter. Recoil is irrelevant for optical light but reshapes hard X-ray and gamma-ray spectra photon by photon.

The story of electron scattering runs from the discovery of the electron to the birth of quantum electrodynamics, and it settled a great argument of twentieth-century physics: whether light carries momentum in discrete lumps. J.J. Thomson, who discovered the electron in 1897, worked out the classical, frequency-independent cross section that bears his name. But when Arthur Compton scattered X-rays off graphite in 1923 and found the scattered light reddened by an angle-dependent amount no wave theory could explain, the classical picture broke. Six years later Oskar Klein and Yoshio Nishina used Dirac's new relativistic electron theory to compute the exact cross section at all energies — one of the first non-trivial predictions of quantum electrodynamics.

Texts. Rybicki & Lightman, Radiative Processes in Astrophysics, Ch. 3 (Thomson) and Ch. 7 (Compton, Klein–Nishina, Comptonization) · Carroll & Ostlie, An Introduction to Modern Astrophysics, 2nd ed., §9.2 (opacity) and §5.1 · Dodelson, Modern Cosmology, Ch. 3 (Compton scattering and the Boltzmann equation) · Mo, van den Bosch & White, Galaxy Formation and Evolution, §6.7 (SZ effect) · Ryden, Introduction to Cosmology, 2nd ed., Ch. 8 (recombination) · Jackson, Classical Electrodynamics, 3rd ed., §14.7.

Provenance. The constants ($\sigma_T = 6.6524\times10^{-29}\ \mathrm{m^2}$, $r_e = 2.818\times10^{-15}\ \mathrm{m}$, $\lambda_C = 2.426\times10^{-12}\ \mathrm{m}$) are CODATA values. The Larmor derivation of $\sigma_T$ and the four-momentum derivation of the Compton shift follow standard treatments (Jackson §14.7; Rybicki & Lightman Ch. 3). The Klein–Nishina formula is quoted from Rybicki & Lightman eq. 7.5. The opacity $\kappa_{es} = 0.02(1+X)\ \mathrm{m^2\,kg^{-1}} = 0.2(1+X)\ \mathrm{cm^2\,g^{-1}}$ is from Carroll & Ostlie §9.2. A verifier should recompute: $\sigma_T = (8\pi/3)r_e^2$; the maximum shift $2\lambda_C = 4.85\ \mathrm{pm}$; $\sigma_{KN}/\sigma_T = 0.431$ at $\varepsilon=1$; $L_{Edd} = 4\pi GMc/\kappa_{es} = 1.5\times10^{31}\ \mathrm{W}$ per solar mass at $X=0.7$; and $y = (kT_e/m_ec^2)\,n_e\sigma_T L$.

Prerequisites. Specific Intensity and Radiative Flux · Opacity and Absorption Coefficients · The Equation of Radiative Transfer · Blackbody Radiation and the Planck Law. Leads to. Optical Depth · Emission and the Source Function · Kirchhoff's Laws and Thermal Emission.

Ask the tutor: Want to see how the same $\sigma_T$ that floors a star's opacity also sets the Eddington limit, polarizes the CMB, and reddens a quasar's X-rays — or work a Comptonization spectrum step by step? Ask away.