Bremsstrahlung & radiation length

A muon and an electron of the same 10 GeV energy enter a block of lead. The muon bores straight through nearly a metre of metal, leaving a thin line of ionisation; the electron is stopped in a few centimetres, dumping its energy into a fizzing cascade of photons, positrons, and more electrons. The two particles differ only in mass, yet their fates in matter could not be more opposite — and the reason is bremsstrahlung, the German for "braking radiation". When a fast electron swerves in the electric field of a nucleus of charge $+Ze$ it accelerates, and an accelerating charge radiates. Above a material-dependent critical energy $E_c$ — a few MeV in lead — this radiative loss overtakes the gentle ionisation loss of the Bethe–Bloch formula and becomes the whole story of how the electron slows down. The rate of radiation is savagely sensitive to mass: it scales as $1/m^2$ , so a muon radiates only $(m_e/m_\mu)^2 \approx 2.3\times10^{-5}$ as fiercely as an electron in the same field. That single factor sorts the particle world into penetrating tracks and showering blobs. And the whole electromagnetic story of high-energy electrons and photons collapses onto one length — the radiation length $X_0$ — the distance over which an electron's energy falls by $1/e$ , and, to within a factor of $7/9$ , the mean free path of a photon before it makes an $e^+e^-$ pair. Because electrons radiate photons and photons make electrons, $X_0$ is the yardstick of the electromagnetic shower, and the design ruler of every calorimeter ever built.

Key takeaways

Above a critical energy $E_c$ (a few MeV in lead), electrons lose energy chiefly by bremsstrahlung — radiating photons in nuclear fields — rather than by ionisation.
The radiative rate scales as $1/m^2$ , suppressed by $(m_e/m_\mu)^2 \approx 2.3\times10^{-5}$ for muons: this is why muons penetrate and electrons shower.
The radiation length $X_0$ is the mean distance over which an electron's energy drops by $1/e$ : $\langle E\rangle = E_0\,e^{-x/X_0}$ .
$X_0$ shortens for high- $Z$ materials as $\approx A/[Z(Z+1)]$ : $X_0(\mathrm{Pb}) = 0.56$ cm, $X_0(\mathrm{Fe}) = 1.76$ cm, $X_0(\mathrm{H_2O}) = 36$ cm.
A high-energy photon pair-produces after a mean free path of exactly $\tfrac{9}{7}X_0$ , so bremsstrahlung and pair production feed one another into an electromagnetic shower.
Key idea: a high-energy electron does not gently ionise its way to rest — it radiates photons in nuclear fields, and the radiation length $X_0$ is the single yardstick for how fast it does so and how the resulting electromagnetic shower grows.

The core insight

A high-energy electron does not gently lose energy by ionisation; it radiates photons in nuclear fields, and the radiation length $X_0$ is the single yardstick measuring how quickly it does so and how an electromagnetic shower develops.

Where this lives today

Radiation length is the master ruler of collider calorimetry. The CMS electromagnetic calorimeter is built from lead-tungstate ( $\mathrm{PbWO_4}$ ) crystals with $X_0 = 0.89$ cm, so its $23$ cm depth packs about $26\,X_0$ — enough to fully contain TeV-scale electron and photon showers in a device you can hold in two hands; ATLAS instead interleaves lead absorbers with liquid-argon sampling gaps to the same end. Distinguishing an electron from a pion, or a photon from a jet, comes down to reading the shower's shape in units of $X_0$ and $R_M$ (the Molière radius, $\approx 1.6$ cm in lead, which sets the transverse spread). Away from colliders, the same physics runs the sky: cosmic-ray air showers develop over an atmospheric $X_0 \approx 300$ m, and experiments like the Pierre Auger Observatory reconstruct $10^{20}$ eV primaries from the shower they seed. Every hospital X-ray tube and every radiotherapy linac is a bremsstrahlung source by design. And at the extreme energies of the highest cosmic rays, an open and actively studied wrinkle appears — the Landau–Pomeranchuk–Migdal (LPM) effect, in which the electron's radiation is coherently suppressed because it takes many atoms' worth of path to form a single soft photon, lengthening the effective radiation length and stretching showers in ways that must be modelled for the next generation of ultra-high-energy detectors.

The physical picture

Bremsstrahlung is the simplest process in electrodynamics dressed up in relativistic clothes: a charge is accelerated and it radiates. What makes it powerful in a detector is what does the accelerating — the enormous Coulomb field close to an atomic nucleus. An electron skimming past a nucleus of charge $+Ze$ feels a transverse kick, its trajectory bends, and the sudden change in velocity shakes a photon loose. The photon carries off a fraction $y = k/E$ of the electron's energy, with the characteristic spectrum favouring soft photons ( $1/k$ ), and it is emitted into a narrow forward cone of half-angle $\theta \sim m_e c^2/E = 1/\gamma$ — for a 10 GeV electron, about $50\ \mu\mathrm{rad}$ . The first figure fixes the geometry and the all-important mass scaling.

Figure

An electron of energy E is deflected in the Coulomb field of a nucleus (+Ze) and radiates a photon of energy k = yE into a forward cone of half-angle θ ≈ 1/γ. The radiated power scales as 1/m², so a muon in the same field radiates only (mₑ/m_μ)² = 2.3×10⁻⁵ as strongly as an electron.

The electron does not radiate once and stop — it radiates again and again, and the photons it makes are themselves unstable in matter: a photon above about 10 MeV converts into an $e^+e^-$ pair in the same nuclear fields (pair production is bremsstrahlung run backwards). Each new electron and positron radiates more photons, each photon makes more pairs, and the number of particles multiplies while the energy per particle falls. This runaway is the electromagnetic shower, and it stops only when the particles drop below $E_c$ and start losing energy to ionisation faster than they can radiate. The second figure is the quantity a calorimeter designer actually reaches for: how the surviving electron energy — and hence the shower it seeds — is metered out in units of $X_0$ .

Figure

The mean energy of an electron falls exponentially with depth measured in radiation lengths, ⟨E⟩ = E₀e^(−x/X₀), dropping to 1/e = 0.368 of its start after exactly one X₀. One radiation length is 0.56 cm in lead, 1.76 cm in iron, and 36 cm in water — the same curve, three rulers.

The core equations

The radiation length is defined by making the radiative energy loss proportional to the energy itself. Unlike ionisation, which removes a roughly fixed energy per unit length, bremsstrahlung removes a fixed fraction per unit length — and a fixed fractional loss integrates to an exponential.

-\frac{dE}{dx}\bigg|_{\text{rad}} = \frac{E}{X_0} \qquad\Longrightarrow\qquad \langle E(x)\rangle = E_0\,e^{-x/X_0} \qquad (1)

where: $E$ — electron energy; $x$ — path length in the material; $X_0$ — radiation length (a property of the material, not the electron); $E_0$ — initial energy. Reading it: after one $X_0$ the mean energy is $E_0/e$ ; after $t = x/X_0$ radiation lengths it is $E_0 e^{-t}$ . The average is what matters — a single electron loses its energy in a few big radiative gulps, not smoothly. The radiation length is computed from the Bethe–Heitler bremsstrahlung cross section. The full Tsai formula is standard; the practical approximation below is accurate to a few percent for $Z > 4$ and is the working tool for estimates.

\frac{1}{X_0} = 4\alpha\,r_e^2\,\frac{N_A}{A}\,Z(Z+1)\ln\!\frac{287}{\sqrt{Z}} \qquad\Longrightarrow\qquad X_0 \approx \frac{716.4\,A}{Z(Z+1)\ln(287/\sqrt{Z})}\ \ \mathrm{g\,cm^{-2}} \qquad (2)

where: $\alpha = 1/137$ — fine-structure constant; $r_e = 2.82\ \mathrm{fm}$ — classical electron radius; $N_A$ — Avogadro's number; $A$ — atomic mass number; $Z$ — nuclear charge. The $Z^2$ (now $Z(Z+1)$ , the extra $Z$ from the atomic electrons) is why high- $Z$ absorbers are compact, and the logarithm is the screening factor $L_{\text{rad}} \approx \ln(183/Z^{1/3})$ . Expressing $X_0$ in $\mathrm{g\,cm^{-2}}$ removes the density and makes it a nearly material-intrinsic number; divide by $\rho$ to get centimetres. The critical energy $E_c$ is the crossover: below it ionisation wins, above it radiation wins. It is set by the material and marks where an electron stops penetrating and starts showering.

E_c \approx \frac{610\ \mathrm{MeV}}{Z+1.24}\ \ \text{(solids, liquids)}, \qquad \frac{(dE/dx)_{\text{rad}}}{(dE/dx)_{\text{ion}}} \approx \frac{E}{E_c} \approx \frac{E\,Z}{800\ \mathrm{MeV}} \qquad (3)

where: $E_c$ — critical energy; the ratio on the right compares radiative to ionisation loss and equals one by definition at $E = E_c$ . The rough form $E_c \approx 800/Z$ MeV follows from the $EZ/800$ estimate; the fitted $610/(Z+1.24)$ is more accurate. For lead $E_c \approx 7.3$ MeV, for iron $\approx 22$ MeV, for air $\approx 87$ MeV. Finally, the shape of the radiated spectrum and the link to pair production. The photon spectrum is soft (roughly $1/k$ ), and the photon's own fate closes the loop into a shower.

\frac{d\sigma}{dk} = \frac{A}{X_0\,N_A}\,\frac{1}{k}\left(\frac{4}{3} - \frac{4}{3}y + y^2\right),\quad y = \frac{k}{E}, \qquad \lambda_{\text{pair}} = \frac{9}{7}\,X_0 \qquad (4)

where: $k$ — radiated photon energy; $y = k/E$ — energy fraction; $\lambda_{\text{pair}}$ — mean free path for a high-energy photon to convert to $e^+e^-$ . Reading it: the $1/k$ makes soft photons overwhelmingly likely (though each carries little energy), and the photon converts after $\tfrac{9}{7}X_0 \approx 1.29\,X_0$ , so the same $X_0$ governs both halves of the cascade.

Derivation ARadiation length and the exponential fall of energy

The goal is to show why the radiative loss takes the clean form $-dE/dx = E/X_0$ , why that integrates to a pure exponential in depth, and why the same length $X_0$ that governs the electron's energy loss also governs — up to the factor $7/9$ — the photon's conversion into pairs. This is the derivation that turns a messy quantum cross section into a single material constant and one exponential law, and it is what lets an experimentalist quote "twenty-five radiation lengths" and know the shower is contained.

Step 1. Start from the Bethe–Heitler cross section for emitting a photon of energy $k$ . In the high-energy, complete-screening limit its energy-weighted integral $\int_0^E k\,(d\sigma/dk)\,dk$ is proportional to $E$ itself, because the $1/k$ spectrum of equation (4) makes $k\,d\sigma/dk$ nearly flat in $k$ and the upper limit scales with $E$ . Why: a flat-in- $k$ spectrum up to $k=E$ radiates a fixed fraction of $E$ per collision, so the mean energy lost per unit length is proportional to $E$ , not constant like ionisation.

Step 2. Multiply by the number density of nuclei $n = N_A\rho/A$ to get the loss per unit length, and define the constant of proportionality to be $1/X_0$ : $-dE/dx|_{\text{rad}} = E/X_0$ . Collecting the atomic factors reproduces equation (2), $1/X_0 \propto (N_A/A)Z^2 r_e^2 \ln(\cdots)$ . Why: bundling every material- and energy-independent factor into one symbol $X_0$ is the whole trick — it hides the quantum mechanics inside a measured length.

Step 3. Integrate the separable equation $dE/E = -dx/X_0$ from $E_0$ to $E$ and $0$ to $x$ : $\ln(E/E_0) = -x/X_0$ , hence $\langle E(x)\rangle = E_0 e^{-x/X_0}$ . After $x = X_0$ the energy is $E_0/e$ . Why: a constant fractional loss rate is the definition of exponential decay — this is the same mathematics as radioactive decay, with depth playing the role of time.

Step 4. Note that the photon that was radiated undergoes pair production, whose cross section in the same screening limit is $\tfrac{7}{9}$ of the brem-equivalent, giving a conversion mean free path $\lambda_{\text{pair}} = \tfrac{9}{7}X_0$ . Why: the $7/9$ is not a coincidence — bremsstrahlung and pair production are the same Feynman diagram with an incoming and outgoing line swapped, so their rates are locked together and one length describes both, which is exactly why the shower is self-similar in units of $X_0$ .

Go deeper · Derivation B · 1 minWhy radiation scales as one over mass squared

Derivation BWhy radiation scales as one over mass squared

The aim is to derive the single most consequential fact about bremsstrahlung — that its rate falls as $1/m^2$ — from elementary classical electrodynamics, and to use it to explain the roughly $10^5$ -fold gulf between how electrons and muons behave in a detector. This is a short argument with an enormous payoff: it is the reason muon spectrometers sit on the outside of every collider experiment and electrons are caught in the calorimeter within.

Step 1. Larmor's formula gives the instantaneous radiated power of a non-relativistic charge as $P = \tfrac{2}{3}\,\dfrac{e^2 a^2}{4\pi\varepsilon_0 c^3}$ , so the radiated power is proportional to the square of the acceleration $a$ . Why: radiation is sourced by acceleration, and the intensity goes as the acceleration squared — a gently deflected charge barely radiates, a violently kicked one blazes.

Step 2. In a given nuclear field the force is $F = Ze^2/(4\pi\varepsilon_0 r^2)$ , fixed by the charges alone, so the acceleration a particle actually undergoes is $a = F/m \propto 1/m$ . Why: the field sets the force, but Newton's second law says the lighter particle whips around faster — the mass enters here, and only here.

Step 3. Combine: $P \propto a^2 \propto 1/m^2$ . Quantum-mechanically the same scaling appears through the cross section $\sigma \propto r_e^2 = \big(e^2/4\pi\varepsilon_0 m_e c^2\big)^2 \propto 1/m^2$ , since the classical radius $r_e$ carries the mass in its denominator. Why: whether you argue classically (Larmor) or with the Bethe–Heitler amplitude (the $r_e^2$ prefactor), the mass appears squared in the denominator — the two pictures agree.

Step 4. Evaluate the muon suppression: $(m_e/m_\mu)^2 = (0.511/105.7)^2 = 2.34\times10^{-5}$ . A muon therefore has a radiation length $\sim 4\times10^4$ times longer than an electron of the same energy, so muon bremsstrahlung only becomes important above a muon critical energy of hundreds of GeV (about 141 GeV in lead). Why: this one number is why a 10 GeV muon punches through a metre of iron while a 10 GeV electron showers away in a few centimetres — same energy, same field, mass squared.

Go deeper · Derivation C · 1 minThe Heitler model of an electromagnetic shower

Derivation CThe Heitler model of an electromagnetic shower

The purpose here is to build the crudest possible model of an electromagnetic shower — a binary tree in which energy halves and particle number doubles every radiation length — and to extract from it the two numbers a calorimeter is designed around: the depth of maximum development $t_{\max}$ and the number of particles there, $N_{\max}$ . Despite its brutality the model gets the scaling laws exactly right, which is why it is the first thing every detector physicist reaches for.

Step 1. Idealise the cascade: after each radiation length every electron radiates one photon carrying half its energy, and every photon converts to an $e^+e^-$ pair splitting its energy in two. Then at depth $t = x/X_0$ there are $N(t) = 2^t$ particles, each with energy $E(t) = E_0/2^t$ . Why: pinning both the brem and pair steps to one radiation length and a clean 50/50 split turns the shower into a doubling tree, which is analytically transparent even though nature is messier.

Step 2. The multiplication cannot continue forever: it halts when the per-particle energy falls to the critical energy $E_c$ , below which ionisation drains energy faster than radiation can multiply particles. Set $E(t_{\max}) = E_c$ : $E_0/2^{t_{\max}} = E_c$ . Why: $E_c$ is the natural stopping point because below it the "double every $X_0$ " engine loses to ionisation — the shower stops growing and starts dying.

Step 3. Solve for the depth of shower maximum: $t_{\max} = \dfrac{\ln(E_0/E_c)}{\ln 2} = \log_2\!\frac{E_0}{E_c}$ , which grows only logarithmically with energy. Why: the logarithm is the headline result — a calorimeter deep enough for 10 GeV needs only a few more radiation lengths for 1 TeV, so detectors stay compact even at collider energies.

Step 4. The particle count at maximum is $N_{\max} = 2^{t_{\max}} = E_0/E_c$ , growing linearly with energy, and the total track length (hence total ionisation signal) is proportional to $E_0/E_c$ . Why: because the deposited signal is proportional to $E_0$ , an electromagnetic calorimeter measures energy by counting shower particles — and its resolution improves as $1/\sqrt{E_0}$ from the Poisson statistics of that count.

Limiting cases

Case	Result	Meaning
$E \gg E_c$ (multi-GeV $e^-$ )	radiation dominates; $\langle E\rangle = E_0 e^{-x/X_0}$	the electron showers; this is the collider regime
$E \ll E_c$ (low-energy $e^-$ )	ionisation dominates (Bethe–Bloch)	slow electrons behave like any charged particle
Heavy particle (muon, proton)	brem suppressed by $(m_e/m)^2$	muons penetrate; brem matters only above $\sim$ hundreds of GeV
High $Z$ absorber	$X_0 \propto A/[Z(Z+1)]$ , small	lead, tungsten give compact calorimeters
Complete screening (high $E$ )	cross section energy-independent; $X_0$ constant	radiation length does not depend on energy
Photon in matter	$\lambda_{\text{pair}} = \tfrac{9}{7}X_0$	photons convert; the shower is self-sustaining
Ultra-high $E$ / dense medium	LPM effect: $X_0$ effectively lengthens	radiation suppressed as amplitudes from many atoms interfere

Common trapBremsstrahlung matters equally for every charged particle

It does not — and the error is a factor of tens of thousands. Because the radiative rate scales as $1/m^2$ , bremsstrahlung is overwhelmingly an electron and positron phenomenon at ordinary energies. For a muon the rate is down by $(m_e/m_\mu)^2 = 2.3\times10^{-5}$ ; for a proton by $(m_e/m_p)^2 \approx 3\times10^{-7}$ . This is precisely why muons are the penetrating particles that reach the outermost detectors while electrons are absorbed in the calorimeter, and why " $dE/dx$ " for a muon in the Bethe–Bloch sense stays valid to very high energy. Bremsstrahlung does eventually catch up even for muons — its loss grows linearly with energy while ionisation grows only logarithmically — but the crossover (the muon critical energy) is at a few hundred GeV in dense materials, not a few MeV. A second, subtler trap: $X_0$ is a property of the material, not of the particle's energy — the same $X_0$ describes a 1 GeV and a 1 TeV electron, because in the complete-screening limit the cross section is energy-independent.

Worked examples

1 · The radiation length of lead (fully worked)

Estimate $X_0$ for lead ( $Z = 82$ , $A = 207.2\ \mathrm{g\,mol^{-1}}$ , $\rho = 11.35\ \mathrm{g\,cm^{-3}}$ ) using the practical formula, and convert to a physical thickness.

Step 1. Assemble the pieces of equation (2). The nuclear factor is $Z(Z+1) = 82\times83 = 6806$ . The screening logarithm needs $\sqrt{Z} = \sqrt{82} = 9.055$ , so $287/\sqrt{Z} = 31.69$ and $\ln(31.69) = 3.456$ .

Step 2. Numerator and denominator: $716.4\times A = 716.4\times 207.2 = 1.484\times10^{5}$ , and the denominator is $Z(Z+1)\ln(287/\sqrt Z) = 6806\times 3.456 = 2.353\times10^{4}$ . Divide: $X_0 = 1.484\times10^{5}/2.353\times10^{4} = \mathbf{6.31\ g\,cm^{-2}}$ (the full Tsai formula gives $6.37$ , so the approximation is good to $1\%$ ).

Step 3. Convert to length by dividing out the density: $X_0 = 6.31/11.35 = \mathbf{0.556\ cm}$ . So a photon or electron shower in lead develops on the scale of half a centimetre per radiation length — the reason lead-based calorimeters are so compact. X₀(Pb) ≈ 6.3 g cm⁻² = 0.56 cm.

2 · Why a muon does not shower (the mass contrast)

A 10 GeV electron and a 10 GeV muon each enter lead. Compare their fates.

Step 1. For the electron, $E = 10\ \mathrm{GeV} \gg E_c(\mathrm{Pb}) = 7.3\ \mathrm{MeV}$ , so radiation dominates by the ratio $E/E_c = 10^{4}/7.3 \approx \mathbf{1400}$ : the electron loses energy almost entirely by bremsstrahlung and showers within a handful of $X_0$ , i.e. a few centimetres of lead.

Step 2. For the muon, the radiative rate is suppressed by $(m_e/m_\mu)^2 = 2.3\times10^{-5}$ , so its effective radiation length is $\sim 4\times10^4$ times longer than the electron's. Equivalently, the muon critical energy in lead is about $\mathbf{141\ GeV}$ , far above 10 GeV, so the muon is nowhere near its radiative regime.

Step 3. The 10 GeV muon therefore loses energy almost entirely by ionisation as a minimum-ionising particle, roughly $12.8\ \mathrm{MeV\,cm^{-1}}$ in lead, and travels of order a metre before stopping — punching cleanly through the calorimeter to the muon chambers outside. Same energy, same field: the electron showers in centimetres, the muon penetrates a metre — the whole difference is mass squared.

3 · Depth of an electromagnetic shower (100 GeV in lead)

A 100 GeV electron enters a lead calorimeter. Where does the shower peak, how many particles are there, and how deep must the detector be?

Step 1. Use the Heitler model. The energy ratio is $E_0/E_c = 10^{5}\ \mathrm{MeV}/7.3\ \mathrm{MeV} = 1.37\times10^{4}$ . The depth of maximum is $t_{\max} = \log_2(E_0/E_c) = \ln(1.37\times10^4)/\ln 2 = 9.53/0.693 = \mathbf{13.7\ X_0}$ .

Step 2. In real showers the profile is a bit shallower than the doubling tree predicts; the measured shower maximum follows $t_{\max} \approx \ln(E_0/E_c) - 0.5 \approx \mathbf{9.0\ X_0}$ (the Heitler model overshoots the depth because it ignores that photons carry energy deeper than electrons). Either way, in lead ( $X_0 = 0.56$ cm) the shower peaks at roughly $9\times0.56 \approx \mathbf{5\ cm}$ .

Step 3. The particle count at maximum is $N_{\max} = E_0/E_c = \mathbf{1.4\times10^{4}}$ — that many electrons, positrons, and photons at the shower's height. To contain $95\%$ of a 100 GeV shower you need roughly $t_{\max} + 0.08\,Z + 9.6 \approx 22$ radiation lengths, i.e. about $\mathbf{12\ cm\ of\ lead}$ . Peak near 9 X₀ ≈ 5 cm, ~1.4×10⁴ particles, full containment in ~22 X₀ ≈ 12 cm.

Physics of the portrait

The quantitative theory of bremsstrahlung and pair production — and with it the very idea of a radiation length — arrived in a single 1934 paper by two young theorists working out the consequences of Dirac's brand-new relativistic electron theory.

Origins · 1 minHans Bethe and how this was discovered

Hans Bethe

1906–2005 · quantum theory of bremsstrahlung and pair production

Hans Bethe was the twentieth century's supreme calculator of how matter and radiation interact, and bremsstrahlung was one of his first great targets. In 1934, with Walter Heitler, he applied Dirac's relativistic quantum mechanics and the Born approximation to an electron scattering in a nuclear Coulomb field and derived the cross sections for both bremsstrahlung and its crossed partner, pair production — the Bethe–Heitler formula that still defines the radiation length today. It was of a piece with his legendary thoroughness: the same Bethe wrote the 1930 stopping-power theory behind the Bethe–Bloch formula (ionisation, the process bremsstrahlung competes with), the three-part 1936–37 "Bethe Bible" that organised all of nuclear physics, and in 1938 the carbon–nitrogen–oxygen cycle that powers the stars, for which he received the 1967 Nobel Prize. During the war he led the Theoretical Division at Los Alamos. What ties the strands together is a physicist who insisted on turning principles into numbers you could check against an experiment — and the radiation length, a single length that collapses a forbidding quantum cross section into a ruler an engineer can lay against a block of lead, is Bethe's method distilled. Walter Heitler, his co-author, went on to write The Quantum Theory of Radiation, the text that taught the subject to a generation.

History

1895

— Röntgen produces X-rays as electrons brake in an anode, bremsstrahlung before it is named

1909

— Sommerfeld develops the classical theory of the continuous X-ray spectrum

1934

— Bethe and Heitler derive the quantum cross sections for bremsstrahlung and pair production

1937

— Carlson and Oppenheimer, and Bhabha and Heitler, formulate electromagnetic cascade theory

1941

— Rossi and Greisen tabulate the radiation length and cascade-shower functions

1974

— Tsai publishes the complete radiation-length formula still used by the Particle Data Group

Going further

Texts. Thomson, Modern Particle Physics, §1.2.2 (interactions of electrons and photons) and §1.2.1 (ionisation, Eq. 1.1–1.2) · Particle Data Group, Review of Particle Physics, "Passage of Particles Through Matter" · Perkins, Introduction to High Energy Physics, Ch. 2 · Leo, Techniques for Nuclear and Particle Physics Experiments, Ch. 2 · Rossi, High-Energy Particles (1952), the classic on radiation length and showers. The key numerical benchmarks are: $X_0(\mathrm{Pb}) = 6.3\ \mathrm{g\,cm^{-2}} = 0.56$ cm and $X_0(\mathrm{Fe}) = 1.76$ cm from equation (2); $(m_e/m_\mu)^2 = 2.34\times10^{-5}$ ; $E_c(\mathrm{Pb}) = 610/(82+1.24) = 7.3$ MeV; for a 100 GeV shower, $t_{\max} = \log_2(10^5/7.3) = 13.7\,X_0$ and $N_{\max} = E_0/E_c = 1.4\times10^4$ ; and $\lambda_{\text{pair}} = \tfrac{9}{7}X_0$ . The shower scaling uses the Heitler doubling model with the Rossi–Greisen longitudinal-profile correction $t_{\max} \approx \ln(E_0/E_c) + C$ .

Prerequisites. Ionisation energy loss (Bethe–Bloch) · Particle–matter interaction · Natural units. Leads to. Electromagnetic showers · Collider experiments and calorimetry.

Ask the tutor: Ask me to walk through why a 10 GeV electron showers in a few centimetres of lead while a 10 GeV muon sails through a metre — the whole answer is one factor of mass squared.

Why does the radiation rate go as one over mass squared?

Because radiation is powered by acceleration, and acceleration is force divided by mass. In a given nuclear field the force on a passing charge is fixed, so a lighter particle accelerates harder — its acceleration goes as $1/m$ . Larmor's formula says radiated power goes as acceleration squared, hence as $1/m^2$ . Quantum-mechanically the same $1/m^2$ shows up because the bremsstrahlung cross section carries a factor of the classical radius squared, $r_e^2$ , and $r_e \propto 1/m$ . The upshot is a factor of $(m_\mu/m_e)^2 \approx 43{,}000$ between a muon and an electron — the single most important number for understanding why detectors are laid out the way they are.

Is the radiation length a property of the material or of the particle?

Purely of the material — and, strictly, of the material for electrons and photons. In the high-energy limit the bremsstrahlung cross section becomes independent of energy (complete screening of the nucleus by atomic electrons), so the same $X_0$ describes a 1 GeV electron and a 1 TeV electron in the same lead. It is not a good ruler for muons or protons, whose radiation is suppressed by $1/m^2$ ; for them one uses their own, far longer, effective radiation lengths. So "10 radiation lengths of lead" is an unambiguous statement about electrons and photons, and a nearly meaningless one about muons.

Why is the radiation length usually quoted in g cm⁻² rather than cm?

Because $\mathrm{g\,cm^{-2}}$ (mass per unit area, i.e. thickness times density) removes almost all of the material dependence. The physics of bremsstrahlung depends on how many nuclei the electron passes, which is set by the areal density of atoms, not by geometric thickness. Two materials with very different densities but similar $Z$ have similar $X_0$ in $\mathrm{g\,cm^{-2}}$ , even though their $X_0$ in centimetres differ wildly (lead $0.56$ cm versus water $36$ cm). To get a physical thickness you divide the $\mathrm{g\,cm^{-2}}$ value by the density — which is exactly what turns $6.3\ \mathrm{g\,cm^{-2}}$ into $0.56$ cm for lead.

Where does the exact 9/7 factor between radiation and pair-production come from?

Bremsstrahlung ( $e \to e\gamma$ in a nuclear field) and pair production ( $\gamma \to e^+e^-$ in a nuclear field) are described by the same Feynman diagram with an external line crossed from incoming to outgoing. Because they share the same underlying amplitude, their cross sections in the complete-screening limit are locked in a fixed ratio, and working through the phase-space integrals gives the pair-production cross section as exactly $\tfrac{7}{9}$ of the quantity that defines $1/X_0$ . So the photon's mean free path is $\tfrac{9}{7}X_0 \approx 1.29\,X_0$ . This tight coupling is why an electromagnetic shower is self-similar in units of $X_0$ : both of its alternating steps are governed by the same length.

Does bremsstrahlung ever matter for muons or protons?

Yes, but only at very high energy. Radiative losses grow linearly with energy while ionisation grows only logarithmically, so every charged particle eventually reaches a critical energy above which it radiates. For muons that crossover is a few hundred GeV in dense materials (about 141 GeV in lead, 347 GeV in iron), so TeV-scale muons — as produced at the LHC or in high-energy cosmic rays — do begin to radiate significantly, via bremsstrahlung, pair production, and photonuclear interactions. This matters for measuring the momentum of very energetic muons, which no longer lose energy at the tidy minimum-ionising rate. For protons the mass is larger still and electromagnetic radiation is negligible at any accessible energy; their energy loss is dominated instead by nuclear interactions.

How deep and wide does an electromagnetic shower actually get?

Longitudinally, the shower peaks at a depth that grows only logarithmically with energy — roughly $t_{\max} \approx \ln(E_0/E_c)$ radiation lengths — and $95\%$ containment needs about $t_{\max} + 10\,X_0$ , so of order $20$ – $25\,X_0$ handles everything from a few GeV to a TeV. That is why calorimeters stay compact. Transversely the shower is narrow: about $90\%$ of its energy falls within one Molière radius $R_M \approx 21\ \mathrm{MeV}\times X_0/E_c$ (roughly $1.6$ cm in lead), which is set by multiple scattering of the low-energy electrons rather than by the forward-peaked bremsstrahlung itself. The logarithmic depth and the fixed transverse scale together are what make electron and photon showers so clean to identify.

Why are calorimeters built from lead or tungsten specifically?

Because $X_0 \propto A/[Z(Z+1)]$ , so a high- $Z$ material has a short radiation length and a small critical energy — the shower is both compact and fully developed in a modest thickness. Lead ( $Z = 82$ , $X_0 = 0.56$ cm) and tungsten ( $Z = 74$ , $X_0 = 0.35$ cm) let you pack $25$ radiation lengths into $10$ – $20$ cm, keeping the detector small, cheap, and fast. The low critical energy also means more shower particles per GeV ( $N_{\max} = E_0/E_c$ ), which improves the energy resolution. The trade-off is that a very dense absorber worsens the sampling in a sampling calorimeter, which is why some detectors instead use homogeneous high- $Z$ crystals like $\mathrm{PbWO_4}$ or $\mathrm{BGO}$ that are both the absorber and the light-producing medium.