Resonant Absorption

Absorption is the elementary act by which matter takes energy from light. A two-level atom in its lower state $|1\rangle$ can swallow one photon of frequency $\omega$ and jump to $|2\rangle$ if $\hbar\omega$ matches the level gap $E_2-E_1=\hbar\omega_0$ . The word that matters is resonant: far from $\omega_0$ the atom barely responds, but tuned to the transition its cross section for stopping a photon becomes enormous — for a strong optical line it reaches $\sigma_0\approx1.7\times10^{-13}\ \mathrm{m}^2$ , roughly ten million times the atom's geometric size. That resonant enhancement is why a wisp of sodium vapour paints a black line across the solar spectrum, and why an atomic clock can lock a laser to a single transition. We take the microscopic transition rate from quantum mechanics and extend it to the optics of absorbing media. aq-qm-fermi-golden-rule derives the rate at which the electric-dipole coupling drives the $|1\rangle\!\to\!|2\rangle$ transition, and aq-qm-absorption-stimulated-emission shows that this rate shares its Einstein coefficient with stimulated emission. What optics owns is the resonant cross section itself — the maximum possible value $\sigma_0=3\lambda^2/2\pi$ , the frequency-integrated strength fixed by the oscillator strength, and how these combine into Beer's law for a real beam through a real gas. Here we assemble that picture: the peak resonant cross section (dipole-independent, a pure $\lambda^2$ ), the integrated cross section $\int\sigma\,\mathrm{d}\nu=\pi r_e c f$ that ties absorption strength to the oscillator strength $f$ and — through the Thomas–Reiche–Kuhn sum rule $\sum f=1$ — to the number of electrons, and Beer's law $I=I_0e^{-\sigma N L}$ extended from its general form (aq-optics-absorption-coefficient-beer-lambert) onto a single sharp line. Master these and you can read a spectral line: its depth gives a column density, its area gives an oscillator strength, and its shape gives a temperature.

Key takeaways

Resonant absorption removes exactly one photon from the field as an atom is promoted from $|1\rangle$ to $|2\rangle$ , and is enormously enhanced when the photon energy $\hbar\omega$ matches the transition energy $E_2-E_1$ .
The absorption rate per atom is $B_{12}\rho(\omega)$ (Einstein), proportional to the spectral energy density; microscopically it is the golden-rule rate $\propto|\langle2|\mathbf{d}|1\rangle|^2$ , obeying the dipole selection rules.
The maximum resonant cross section of a two-level atom is a universal $\sigma_0=3\lambda^2/2\pi$ — independent of the dipole strength — reaching $1.66\times10^{-13}\ \mathrm{m}^2$ for the Na D line at 589 nm.
The frequency-integrated cross section is $\int\sigma\,\mathrm{d}\nu=\pi r_e c\,f=2.65\times10^{-6}\,f\ \mathrm{m^2\,Hz}$ , fixed by the oscillator strength $f$ and independent of broadening — Doppler and collisions reshape the line but conserve its area.
The oscillator strengths of a given atom obey the Thomas–Reiche–Kuhn sum rule $\sum_k f_k=Z$ (=1 for a one-electron atom); hydrogen's Lyman-α alone carries $f=0.4162$ .
Macroscopically a beam obeys Beer's law $I(z)=I_0e^{-\alpha z}$ with $\alpha=(N_1-\tfrac{g_1}{g_2}N_2)\sigma$ ; net absorption needs more atoms below than above, and at high intensity the line saturates.
Key idea: resonant absorption removes one photon per upward jump; its peak cross section is a dipole-independent $3\lambda^2/2\pi$ , its frequency-integrated strength is fixed by the oscillator strength ( $\int\sigma\,\mathrm{d}\nu=\pi r_e c f$ , broadening-independent), and its macroscopic law is Beer's $I=I_0e^{-\alpha z}$ .

The core insight

Resonant absorption removes exactly one photon per upward transition; its peak cross section is a universal $3\lambda^2/2\pi$ , its frequency-integrated strength $\int\sigma\,\mathrm{d}\nu=\pi r_e c f$ is fixed by the oscillator strength and independent of the broadening mechanism, and its macroscopic manifestation is Beer's law $I(z)=I_0e^{-\alpha z}$ .

Where this lives today

Resonant absorption is the workhorse of quantitative spectroscopy. Because the frequency-integrated cross section is broadening-independent, the equivalent width of an absorption line yields the product $N_1 f$ directly — the method that measures column densities of elements from laboratory vapour cells to interstellar clouds and quasar absorption systems billions of light-years away. Solar and stellar Fraunhofer lines encode temperature (Doppler width), pressure (collisional width), abundance (line depth), and velocity fields (line shift), making resonant absorption the primary diagnostic of remote and inaccessible matter. The same physics underlies atomic clocks and laser cooling. Near-resonant absorption followed by spontaneous emission transfers momentum $\hbar k$ per cycle; the velocity dependence of the absorption through the Doppler shift produces the friction of optical molasses and magneto-optical traps, cooling atoms to microkelvin temperatures. Atomic-vapour cells (rubidium, cesium) serve as narrowband optical filters and frequency references, and saturable absorbers — media whose absorption bleaches at high intensity — passively mode-lock lasers to make femtosecond pulses. Every one of these devices exploits the enormous, universal resonant cross section $3\lambda^2/2\pi$ that a single atomic transition can present.

The physical picture

Picture the atom as a charge on a spring driven by the oscillating electric field. Off resonance the charge barely moves and takes little energy; tuned to its natural frequency $\omega_0$ the driven oscillation grows and the atom absorbs efficiently. Quantum mechanics replaces the classical energy transfer with a discrete event — one photon leaves the mode, the atom jumps to $|2\rangle$ — but keeps the resonance: the response is sharply peaked at $\omega_0$ and falls away on either side. Absorption is bookkeeping between field and atom: one quantum out of the field, one quantum of internal excitation in. Because the reverse process — stimulated emission — is governed by the same coupling, a beam is attenuated only to the extent that more atoms sit in $|1\rangle$ than in $|2\rangle$ . Equalize the populations and the sample turns transparent; that is the seed of saturation, and the reason absorption and gain are two faces of one interaction. The first figure follows a beam attenuating through a vapour cell; the second is the punchline — the resonant cross section itself, with the universal peak $\sigma_0=3\lambda^2/2\pi$ marked for the Na D line and the integrated strength $\int\sigma\,\mathrm{d}\nu=\pi r_e c f$ that no amount of broadening can change.

Figure

A resonant beam attenuating through a vapour cell of atom density N: Beer's law I(z) = I₀ e^(−σNz). At the Na D resonance the cross section is σ₀ = 3λ²/2π = 1.66×10⁻¹³ m², so a column density N·L = 1/σ₀ = 6.0×10¹² m⁻² already gives optical depth τ = 1 (the beam drops to 1/e)

The resonant cross section is astonishingly large because it is set by the wavelength, not the atomic size. And its frequency-integrated value is even more robust — it is fixed entirely by the oscillator strength and cannot be changed by any broadening mechanism. Doppler motion and collisions widen the line and lower its peak, but they only redistribute a conserved area.

Figure

The resonant absorption line σ(ν): a Lorentzian peaking at σ₀ = 3λ²/2π = 1.66×10⁻¹³ m² (Na D). Broadening (Doppler, collisions) lowers and widens the peak but the shaded area ∫σ dν = π r_e c f = 2.65×10⁻⁶ f m²·Hz is conserved — fixed by the oscillator strength f. Hydrogen's Lyman-α carries f = 0.4162 of the sum-rule total Σf = 1

The core equations

Einstein's phenomenology assigns an absorption rate per lower-state atom proportional to the spectral energy density of the field. This is the macroscopic starting point; the microscopic rate follows from the golden rule.

\left(\frac{\mathrm{d}N_2}{\mathrm{d}t}\right)_{\text{abs}}=B_{12}\,\rho(\omega_0)\,N_1 \qquad (1)

where: $N_1,N_2$ — lower/upper number densities; $B_{12}$ — Einstein absorption coefficient; $\rho(\omega_0)$ — spectral energy density at resonance. Because the rate is linear in $\rho$ , weak fields absorb linearly — the regime of ordinary Beer's law — while strong fields eventually saturate. The microscopic rate comes from time-dependent perturbation theory. aq-qm-fermi-golden-rule gives the transition rate as $2\pi/\hbar$ times the squared matrix element times the density of final states; with the dipole coupling $H'=-\mathbf{d}\cdot\mathbf{E}$ and a line shape $g(\omega)$ replacing the delta function, this becomes an absorption cross section.

\sigma(\omega)=\frac{\pi\omega_0}{\varepsilon_0 c\hbar}\,|\mathbf{d}_{21}|^2\,g(\omega) \qquad (2)

where: $\sigma$ — absorption cross section (area); $\mathbf{d}_{21}=\langle2|e\mathbf{r}|1\rangle$ — transition dipole moment; $g(\omega)$ — normalized line shape, $\int g\,\mathrm{d}\omega=1$ ; $\varepsilon_0$ — vacuum permittivity. The response is proportional to $|\mathbf{d}_{21}|^2$ : a strong dipole absorbs strongly, and the selection rules that make $\mathbf{d}_{21}$ vanish forbid the line entirely. A beam through a gas of absorbers is attenuated exponentially — Beer's law, extended from its general form (recall aq-optics-absorption-coefficient-beer-lambert) onto a resonant line by making $\alpha$ frequency-dependent and population-weighted.

I(z)=I_0\,e^{-\alpha z},\qquad \alpha(\omega)=\left(N_1-\frac{g_1}{g_2}N_2\right)\sigma(\omega) \qquad (3)

where: $\alpha$ — absorption coefficient (per length); $g_1,g_2$ — level degeneracies. The population factor $N_1-(g_1/g_2)N_2$ is the mirror image of the gain condition: net absorption needs more atoms below than above, and equal populations give transparency. The single most useful fact is that the frequency-integrated cross section is fixed by the oscillator strength and independent of the line shape.

\int\sigma(\nu)\,\mathrm{d}\nu=\pi r_e c\,f_{12},\qquad r_e=\frac{e^2}{4\pi\varepsilon_0 m_e c^2} \qquad (4)

where: $f_{12}$ — absorption oscillator strength; $r_e=2.818\times10^{-15}$ m — classical electron radius. Because $\int g\,\mathrm{d}\nu=1$ , broadening redistributes the same total area: the peak drops and the wings fill, but $\int\sigma\,\mathrm{d}\nu$ is conserved. (In angular frequency the same statement reads $\int\sigma\,\mathrm{d}\omega=\pi e^2 f_{12}/2\varepsilon_0 m_e c$ .) At exact resonance, for a purely radiatively (natural-width) broadened transition, the peak cross section reduces to a universal geometric value — the maximum a two-level atom can present.

\sigma_0=\frac{3\lambda^2}{2\pi}\qquad(\text{two-level, on resonance, natural width}) \qquad (5)

where: $\sigma_0$ — peak resonant cross section; $\lambda$ — resonant wavelength in the medium. The dipole moment cancels completely (the same cancellation as the gain cross section $\lambda^2/2\pi$ , here carrying the degeneracy/orientation factor 3). The oscillator strengths of an atom are not free: they obey a sum rule.

\sum_k f_k = Z\quad(\text{Thomas–Reiche–Kuhn}),\qquad f_{12}=\frac{2m_e\omega_0}{3\hbar e^2}|\mathbf{d}_{21}|^2 \qquad (6)

where: $Z$ — number of electrons; $f_k$ — oscillator strength of transition $k$ . For a one-electron atom $\sum f=1$ , and a single strong line (hydrogen Lyman-α, $f=0.4162$ ) can carry a large fraction of the total absorbing capacity.

Derivation AThe maximum resonant cross section 3λ²/2π

The goal is to prove that at exact resonance a radiatively broadened two-level atom presents a peak absorption cross section $\sigma_0=3\lambda^2/2\pi$ that depends only on the wavelength — the dipole moment cancels entirely. The strategy is to insert the Lorentzian line-shape peak $g(\omega_0)=2/\pi\Gamma$ and the natural width $\Gamma=A_{21}$ (expressed through the same dipole moment that appears in $\sigma$ ) into the cross section, and watch $|\mathbf{d}_{21}|^2$ divide out. The one caveat: this ceiling applies only to natural broadening; Doppler or collisional widening pushes the real peak below $3\lambda^2/2\pi$ .

Step 1. Evaluate the cross section at line centre. A normalized Lorentzian of FWHM $\Gamma$ (in angular frequency) peaks at $g(\omega_0)=\dfrac{2}{\pi\Gamma}$ , so $\sigma_0=\dfrac{\pi\omega_0}{\varepsilon_0 c\hbar}|\mathbf{d}_{21}|^2\cdot\dfrac{2}{\pi\Gamma}$ . Why the Lorentzian: the finite radiative lifetime $\tau$ of the upper state gives an exponentially decaying amplitude whose Fourier transform is Lorentzian, with $\Gamma=1/\tau=A_{21}$ — the natural linewidth (owned by aq-qm-lifetime-linewidth-broadening, recalled here by name).

Step 2. Insert the natural width expressed through the dipole moment, $\Gamma=A_{21}=\dfrac{\omega_0^3|\mathbf{d}_{21}|^2}{3\pi\varepsilon_0\hbar c^3}$ . This is the Einstein/Weisskopf spontaneous rate — the same $|\mathbf{d}_{21}|^2$ that sits in the numerator of $\sigma_0$ . Why the dipole appears twice: a strong dipole makes the cross section tall ( $\sigma\propto|\mathbf{d}_{21}|^2$ ) but also makes the natural line wide ( $\Gamma\propto|\mathbf{d}_{21}|^2$ ), and a wider line is shorter — the two effects are destined to cancel.

Step 3. Substitute and cancel: $\sigma_0=\dfrac{2\omega_0|\mathbf{d}_{21}|^2}{\varepsilon_0 c\hbar\,\Gamma}=\dfrac{6\pi c^2}{\omega_0^2}=\dfrac{3\lambda^2}{2\pi}$ , using $\lambda=2\pi c/\omega_0$ . For the Na D line at 589 nm this is $1.66\times10^{-13}\ \mathrm{m}^2$ . Why the factor 3: it is the degeneracy/orientation weight of a dipole transition (for a $J=0\to1$ line, $(2J_2+1)/(2J_1+1)=3$ ); the bare non-degenerate two-level value is $\lambda^2/2\pi$ , the gain cross section recalled from aq-optics-stimulated-emission. Verified in verify/verify_aq-optics-resonant-absorption.py: σ₀ = 3λ²/2π = 1.66×10⁻¹³ m² at 589 nm; dipole cancellation confirmed symbolically; σ₀ = 3 × the non-degenerate λ²/2π.

Go deeper · Derivation B · 1 minThe integrated cross section and the oscillator-strength sum rule

Derivation BThe integrated cross section and the oscillator-strength sum rule

The goal is to show that the frequency-integrated absorption cross section is $\int\sigma\,\mathrm{d}\nu=\pi r_e c\,f_{12}$ — a broadening-independent measure of line strength — and that the oscillator strengths of an atom sum to the number of electrons (Thomas–Reiche–Kuhn). The strategy is to integrate the line-shape-weighted cross section using the normalization $\int g\,\mathrm{d}\nu=1$ , identify the oscillator strength through the dipole moment, and invoke the sum rule as a constraint on the whole spectrum. The payoff is a robust quantitative tool: measure the area of an absorption line and you measure $N f$ directly, immune to how the line is broadened.

Step 1. Integrate the cross section over frequency. Since $\sigma(\nu)=[\text{constant}]\times g(\nu)$ and $\int g\,\mathrm{d}\nu=1$ , the integral is the constant alone: $\int\sigma\,\mathrm{d}\nu=\pi r_e c\,f_{12}$ , with $f_{12}=\dfrac{2m_e\omega_0}{3\hbar e^2}|\mathbf{d}_{21}|^2$ defining the oscillator strength. Why the shape drops out: the line shape only distributes absorption across frequency; its integral is fixed at 1 by normalization, so the total absorbing capacity is set by $f_{12}$ alone and is blind to Doppler or collisional widening.

Step 2. Evaluate the coefficient. $\pi r_e c=\pi(2.818\times10^{-15})(2.998\times10^8)=2.65\times10^{-6}\ \mathrm{m^2\,Hz}$ , so $\int\sigma\,\mathrm{d}\nu=2.65\times10^{-6}f\ \mathrm{m^2\,Hz}$ . Hydrogen's Lyman-α with $f=0.4162$ gives $1.10\times10^{-6}\ \mathrm{m^2\,Hz}$ . Why this is the spectroscopist's workhorse: the observed equivalent width of an absorption line yields $N_1 f_{12}$ directly, so with a known $f$ one reads off the column density $N_1$ — the basis of quantitative absorption spectroscopy from lab cells to quasar sight-lines.

Step 3. Constrain the whole spectrum. Summing the oscillator strengths over all transitions from a given level gives the Thomas–Reiche–Kuhn sum rule $\sum_k f_k=Z$ , the number of electrons — 1 for hydrogen. Thus Lyman-α alone accounts for 42% of hydrogen's entire absorbing capacity from the ground state. Why a sum rule exists at all: it is a completeness statement — the total oscillator strength counts the electrons available to be driven, a consequence of the canonical commutator $[x,p]=i\hbar$ , independent of the atomic potential. Verified in verify/verify_aq-optics-resonant-absorption.py: πr_e c = 2.65×10⁻⁶ m²·Hz; ∫σ dν = πr_e c f; Lyman-α f = 0.4162; Na D f-sum 0.961; TRK Σf = 1.

Limiting cases

Resonant absorption changes character across regimes:

Case	Result	Meaning
Weak field (linear)	$\alpha(\omega)=N_1\sigma(\omega)$ , intensity-independent	Populations barely disturbed; Beer's law holds with fixed $\alpha$ .
Strong field (saturation)	$\sigma\to\sigma_0/(1+I/I_{\text{sat}})$	Stimulated emission competes; the line bleaches as $I\to I_{\text{sat}}$ .
Doppler-dominated gas	$g(\omega)\propto e^{-(\omega-\omega_0)^2/2\sigma_D^2}$	Warm low-pressure vapour: Gaussian profile, width $\propto\omega_0\sqrt{k_BT/mc^2}$ ; peak below $3\lambda^2/2\pi$ .
Natural / collision-limited	Lorentzian, FWHM $\Gamma$	Cold, narrow-line limit: peak reaches $\sigma_0=3\lambda^2/2\pi$ .
Far off resonance	$\sigma\propto1/(\omega-\omega_0)^2\to0$	Absorption vanishes; the response becomes dispersive (refraction) rather than absorptive.

Worked examples

Example 1 — the resonant cross section and column density for optical depth 1 (confidence builder). Compute the peak resonant cross section of the Na D line ( $\lambda=589$ nm, radiatively broadened) and the column density of sodium atoms needed to make a vapour opaque (optical depth $\tau=1$ ).

Step 1. The peak cross section is dipole-independent: $\sigma_0=3\lambda^2/2\pi=3(589\times10^{-9})^2/(2\pi)=\mathbf{1.66\times10^{-13}\ m^2}$ ( $=1.66\times10^{-9}\ \mathrm{cm}^2$ ). This is about $10^7$ times the atom's geometric cross section ( $\pi a_0^2\approx9\times10^{-21}\ \mathrm{m}^2$ ) — the hallmark of resonant response.

Step 2. Optical depth is $\tau=\sigma_0 N L$ ; setting $\tau=1$ requires a column density $NL=1/\sigma_0=1/(1.66\times10^{-13})=\mathbf{6.0\times10^{12}\ m^{-2}}$ ( $=6.0\times10^{8}\ \mathrm{cm}^{-2}$ ). A vapour path with that few atoms per unit area already dims the beam to $1/e$ .

Step 3. A sodium cell 10 cm long thus needs only $N=6.0\times10^{13}\ \mathrm{m}^{-3}$ atoms to reach $\tau=1$ — a vanishingly thin vapour. This extreme sensitivity is why sodium street-lamp light is strongly self-reversed and why the solar sodium line is one of the deepest Fraunhofer lines, even though sodium is a trace constituent of the photosphere. (Real lines are Doppler-broadened, lowering $\sigma_0$ and raising the required column density, but the ordering of magnitudes stands.)

Example 2 — oscillator strength and the sum rule (exam level). Hydrogen's Lyman-α transition (1s→2p, 121.57 nm) has oscillator strength $f=0.4162$ . Find its frequency-integrated cross section and the fraction of hydrogen's total ground-state absorbing capacity it represents.

Step 1. The integrated cross section is $\int\sigma\,\mathrm{d}\nu=\pi r_e c\,f=(2.65\times10^{-6})(0.4162)=\mathbf{1.10\times10^{-6}\ m^2\,Hz}$ , using $\pi r_e c=\pi(2.818\times10^{-15})(2.998\times10^8)=2.65\times10^{-6}\ \mathrm{m^2\,Hz}$ .

Step 2. The Thomas–Reiche–Kuhn sum rule fixes the total: $\sum_k f_k=Z=1$ for hydrogen. Lyman-α therefore carries $f/Z=\mathbf{0.4162}$ — about 42% — of all the oscillator strength available from the ground state, with the rest split among Lyman-β, -γ, … and the continuum.

Step 3. By contrast, sodium's two D lines (D₂ $f=0.641$ , D₁ $f=0.320$ ) sum to $\mathbf{0.961}$ — nearly the entire valence-electron oscillator strength lives in that single doublet. This is why the sodium doublet dominates its absorption and emission spectrum so completely, and why it is the reference transition for atomic clocks, laser cooling, and streetlight spectroscopy.

Physics of the portrait

Absorption — the resonant removal of light by matter — is where spectroscopy was born. In 1814 Joseph von Fraunhofer mapped hundreds of dark lines crossing the solar spectrum without knowing their cause. The explanation came in 1859, when Gustav Kirchhoff and Robert Bunsen showed that a cool gas absorbs exactly the wavelengths it would emit when hot: Kirchhoff's law of thermal radiation tied absorption and emission together and let them identify sodium in the Sun from the D lines. Quantifying "how much" a line absorbs took the concept of oscillator strength, introduced through Rudolf Ladenburg's dispersion work around 1921 and made rigorous by the Thomas–Reiche–Kuhn sum rule in 1925 — a completeness relation that counts electrons. Einstein's 1917 $B$ -coefficient supplied the rate; the golden rule supplied the microscopic mechanism; together they explain both the black Fraunhofer line and the atomic clock.

Origins · 1 minabsorption and how this was discovered

absorption

resonant removal of photons by matter · Fraunhofer 1814 → Kirchhoff–Bunsen 1859 → TRK sum rule 1925 · spectroscopy and light–matter interaction

History

1814

— Fraunhofer catalogues the dark absorption lines in the solar spectrum

1859

— Kirchhoff and Bunsen establish that a gas absorbs the wavelengths it emits, founding spectral analysis

1900

— Planck's radiation law fixes the thermal field that absorption and emission must equilibrate with

1917

— Einstein introduces the absorption coefficient B₁₂ in his rate-equation treatment

1921

— Ladenburg quantifies dispersion and absorption through the oscillator strength

1925

— Thomas, Reiche and Kuhn prove the oscillator-strength sum rule Σf = Z

Going further

References. Foot, Atomic Physics (2005), Ch. 7 (absorption, line shapes) · Loudon, The Quantum Theory of Light, 3rd ed., Ch. 1 & 2 · Saleh & Teich, Fundamentals of Photonics, 2nd ed., Ch. 13–14 · Demtröder, Laser Spectroscopy, Vol. 1, Ch. 2–3 · Milonni & Eberly, Laser Physics, Ch. 3–4 · Rybicki & Lightman, Radiative Processes in Astrophysics, Ch. 10.

Provenance. Equations and numbers are standard textbook results. A verifier recomputes: (1) the detailed-balance relations $g_1B_{12}=g_2B_{21}$ and $A_{21}=\hbar\omega_0^3 B_{21}/\pi^2c^3$ (Loudon Ch. 1; Foot Ch. 7); (2) the golden-rule rate and cross section (Loudon Ch. 2); (3) the universal peak $\sigma_0=3\lambda^2/2\pi$ by cancelling the dipole against $A_{21}$ ; (4) the integrated-cross-section relation $\int\sigma\,\mathrm{d}\nu=\pi r_e c f$ (Foot Ch. 7). The numeric $\sigma_0\approx1.66\times10^{-13}\ \mathrm{m}^2$ at 589 nm follows directly from $3\lambda^2/2\pi$ ; Lyman-α $f=0.4162$ and the Na D f-sum $0.961$ are standard tabulated values.

Builds on: Electric Dipole Approximation · Einstein A and B Coefficients · Selection Rules for Atomic Transitions. Leads to: Stimulated Emission · Detailed Balance and the Planck Spectrum · Oscillator Strength.

How can an atom absorb a photon ten million times bigger than its own cross section?

The cross section is set by the wavelength, not the atomic size. On resonance a two-level atom presents $\sigma_0=3\lambda^2/2\pi$ , which for visible light is $\sim10^{-13}\ \mathrm{m}^2$ — vastly larger than the $\sim10^{-20}\ \mathrm{m}^2$ geometric cross section of the electron cloud. Physically, the atom is a resonant antenna: driven at its natural frequency it "reaches out" a wavelength-scale distance to grab energy from the passing wave, exactly as a resonant radio antenna intercepts far more power than its physical size suggests.

Why doesn't broadening change how much light a line absorbs in total?

Broadening only spreads the absorption over frequency; it cannot create or destroy the total. Mathematically $\sigma(\nu)=[\text{const}]\times g(\nu)$ with $\int g\,\mathrm{d}\nu=1$ , so the integral $\int\sigma\,\mathrm{d}\nu=\pi r_e c f$ is fixed by the oscillator strength alone. Doppler motion and collisions lower the peak and fill the wings, but the area under the line is conserved. That is why spectroscopists measure the equivalent width (the integrated absorption) rather than the peak depth — it is a robust measure of $N f$ regardless of the line shape.

What is the oscillator strength, really?

It is a dimensionless number that says how strongly a transition absorbs compared to a classical electron oscillator. A classical charge on a spring would absorb with $f=1$ ; a real quantum transition absorbs with $f_{12}=2m_e\omega_0|\mathbf{d}_{21}|^2/3\hbar e^2$ , which can be less than 1 (weak) or, summed over all transitions, exactly the number of electrons (Thomas–Reiche–Kuhn). It converts the abstract dipole matrix element into a directly measurable line strength, which is why atomic data tables list $f$ -values rather than dipole moments.

Is resonant absorption the same as stimulated emission run backwards?

Essentially yes. Both are driven by the same first-order dipole coupling and share the same Einstein $B$ -coefficient and the same cross section; the only difference is which level the atom starts in. Absorption takes a lower-state atom up and removes a photon; stimulated emission takes an upper-state atom down and adds one. That symmetry is exactly why a beam is attenuated only when $N_1>(g_1/g_2)N_2$ and amplified when the populations are inverted — absorption and gain are two signs of one coefficient.

Why does the peak cross section not depend on how strong the transition is?

Because the transition's strength sets both the peak height and the natural linewidth in compensating amounts. A large dipole gives a large cross section ( $\sigma\propto|\mathbf{d}_{21}|^2$ ) but also a wide natural line ( $\Gamma\propto|\mathbf{d}_{21}|^2$ ), and a wider line is shorter. The two cancel, leaving $\sigma_0=3\lambda^2/2\pi$ . This universality holds only for natural (radiative) broadening; Doppler and collisional broadening add width without height, so real vapour lines peak below $3\lambda^2/2\pi$ .

What happens to absorption at very high intensity?

It saturates. As the intensity rises toward the saturation intensity $I_{\text{sat}}$ , stimulated emission increasingly competes with absorption and the population difference collapses as $\Delta N=\Delta N_0/(1+I/I_{\text{sat}})$ , so $\alpha\to0$ and the medium bleaches — it becomes transparent to its own resonant light. This saturable behaviour is not a nuisance but a tool: saturable absorbers passively mode-lock lasers, letting only the most intense pulse peaks through with low loss and thereby generating ultrashort pulses.