Hyperfine Structure

Fine structure splits hydrogen's levels by parts in $10^{5}$ ; hyperfine structure splits them again, a thousand times more finely — and that whisper is one of the most useful numbers in physics. Quantum mechanics owns why it exists: aq-qm-hyperfine-structure builds the nuclear-moment coupling $\hat H_{hf}=A\,\hat{\mathbf I}\cdot\hat{\mathbf J}$ , the total-angular-momentum algebra $\mathbf F=\mathbf I+\mathbf J$ , the Landé interval rule, the Fermi contact term, and the Breit–Rabi field dependence. This article recalls that machinery by name and owns its spectroscopic cash value: the evaluation of the hydrogen ground-state splitting to $1420$ MHz, the $21$ -centimetre line that weighs the galaxy, the caesium transition that defines the second, and the beam resonances that read nuclear spin straight off a multiplet. The mechanism is a magnet inside a magnet. The proton is a spin- $\tfrac12$ particle with a magnetic moment; the electron, through its spin and its motion, makes a magnetic field at the nucleus; and the energy of one magnet in another's field depends on their relative orientation. Line the two spins up or oppose them and hydrogen's ground state splits in two, by $5.9$ microelectronvolts. The suppression is a mass ratio, not a power of $\alpha$ : the nuclear magneton is smaller than the Bohr magneton by $m_e/m_p\approx1/1836$ , which is exactly why the effect earns the prefix "hyper." What follows is that one number turned into instruments — a radio line at $21$ cm, a clock ticking $9\,192\,631\,770$ times a second, and a ruler for nuclear spin.

Key takeaways

Hyperfine structure is the nuclear magnetic moment sitting in the electron's field: the dominant coupling is $\hat H_{hf}=A\,\hat{\mathbf I}\cdot\hat{\mathbf J}$ , and the good quantum number becomes $\mathbf F=\mathbf I+\mathbf J$ (nuclear spin now included).
The coupling algebra, interval rule, contact term, and Breit–Rabi formula are owned by aq-qm-hyperfine-structure; here we recall them and evaluate the numbers, the astronomy, and the clock.
Hydrogen's $1s$ ground state splits into $F=1$ (triplet) and $F=0$ (singlet) by $\Delta E=\tfrac43 g_p\tfrac{m_e}{m_p}\alpha^{4}m_ec^{2}=5.87\ \mu$ eV — the $21$ -centimetre line at $1420.405751768$ MHz ( $68.2$ mK, $\lambda=21.106$ cm).
The transition is magnetic-dipole and "forbidden": its rate is $A_{21}\approx2.85\times10^{-15}\ \mathrm{s}^{-1}$ , a lifetime near $11$ million years, yet the sheer abundance of interstellar hydrogen makes it radio astronomy's brightest line.
The suppression is $\mu_N/\mu_B=m_e/m_p\approx1/1836$ — a mass ratio, not another power of $\alpha$ ; nuclei with $I=0$ (¹²C, ¹⁶O) have no hyperfine structure at all.
The caesium $F=3\leftrightarrow F=4$ transition, $9\,192\,631\,770$ Hz, defines the SI second; the $m_F=0\to m_F=0$ pair is chosen because it is field-independent to first order.
Key idea: hyperfine structure is the nucleus written into the atom's angular-momentum bookkeeping — a magnet-in-a-magnet energy, smaller than fine structure by $m_e/m_p$ , whose one hydrogen number became a galactic yardstick and the definition of time.

The core insight

The nucleus is a small magnet, $\hat{\boldsymbol\mu}_I=g_I\mu_N\hat{\mathbf I}/\hbar$ , sitting in the field the electron makes at the origin. Its orientation energy is $\hat H_{hf}=A\,\hat{\mathbf I}\cdot\hat{\mathbf J}/\hbar^{2}$ , so the good quantum number becomes $\mathbf F=\mathbf I+\mathbf J$ and levels split by $\tfrac{A}{2}[F(F+1)-I(I+1)-J(J+1)]$ . The scale is $\mu_N/\mu_B=m_e/m_p\approx1/1836$ below fine structure — hence "hyperfine."

Where this lives today

Hyperfine structure is the physics of the atomic clock. Caesium-133's ground-state transition was defined to be $9\,192\,631\,770$ Hz in 1967, making the second — and through it the metre and the volt — a hyperfine measurement; the interval rule and the field-independent Breit–Rabi transition are what make it reproducible, and cold-atom fountains now realize it to a few parts in $10^{16}$ . Rubidium and hydrogen masers run on the same principle, and optical-lattice clocks push related transitions to $10^{-18}$ . Beyond timekeeping, NMR and MRI read the hyperfine coupling of nuclear spins to their electronic environment, turning chemistry into a spectrum; electron-spin resonance maps unpaired electrons the same way; and hyperfine states of trapped ions and neutral atoms serve as long-lived qubits precisely because the clock transition is field-insensitive. And across the universe the $21$ -cm line — the singlet–triplet transition evaluated above — weighs neutral hydrogen from nearby galaxies back to the epoch of reionization, where its redshifted glow is the target of a new generation of radio arrays.

The physical picture

An electron orbiting a proton is, from the proton's point of view, a ring of current and a nearby bar magnet — the electron's own spin — that together bathe the nucleus in a magnetic field. The proton is itself a magnet, and a magnet in a field has an orientation energy: flip the proton's moment relative to the electron's field and the atom's energy shifts. That shift, resolved by orientation, is hyperfine structure. It is tiny because the nuclear magneton $\mu_N=e\hbar/2m_p$ is smaller than the electronic Bohr magneton by the mass ratio $m_e/m_p$ , and it is special because for an $s$ -state the electron sits literally on top of the point nucleus — the density $|\psi(0)|^{2}$ is nonzero there — so the coupling is the Fermi contact interaction, a delta-function term derived in aq-qm-hyperfine-structure and felt only by states that reach the origin. The coupling knits nuclear spin into the atom's angular-momentum accounting. Before the nucleus mattered the good number was $\mathbf J=\mathbf L+\mathbf S$ ; now $\hat{\mathbf I}\cdot\hat{\mathbf J}$ mixes $m_I$ and $m_J$ , and only their sum survives. The total $\mathbf F=\mathbf I+\mathbf J$ becomes the new good quantum number — the same angular-momentum addition that assembles $\mathbf J$ from $\mathbf L$ and $\mathbf S$ in Fine Structure and Spin-Orbit Coupling, with the nucleus playing the role of a second spin.

Figure

Hydrogen 1s hyperfine splitting: the two spin orientations differ by 5.87 µeV = h·1420.406 MHz = a 21.106 cm photon = 68.2 mK

The interval rule is the fingerprint of magnetic-dipole hyperfine structure: adjacent $F$ levels are spaced $\propto F$ , not equally, so the growing gaps count the $F$ values and pin the nuclear spin. And an external field competes with the internal $\hat{\mathbf I}\cdot\hat{\mathbf J}$ coupling — weak fields keep $F$ good, strong fields tear $\mathbf J$ from $\mathbf I$ — the crossover that runs optical pumping, vapour-cell magnetometers, and atomic clocks. The second figure is why hyperfine structure repays study out of all proportion to its size: the four smallest energy scales of hydrogen span nearly five decades, and the hyperfine rung is the one astronomers listen to.

Figure

The hydrogen energy pyramid (log scale): gross structure 10.2 eV sits about five decades above fine structure 45.3 µeV; hyperfine 5.87 µeV — the 21 cm line — and the Lamb shift 4.37 µeV lie a further decade down

The core equations

The nuclear moment is written in nuclear magnetons, the unit set one mass-ratio below the electron's. This single factor is the entire reason hyperfine structure is small.

\hat{\boldsymbol\mu}_I=g_I\,\mu_N\,\frac{\hat{\mathbf I}}{\hbar},\qquad \mu_N=\frac{e\hbar}{2m_p}=\frac{m_e}{m_p}\,\mu_B \qquad (1)

where: $g_I$ — nuclear $g$ -factor (bare proton $g_p\approx5.586$ ); $\mu_N=5.05\times10^{-27}\ \mathrm{J/T}$ ; the ratio $\mu_N/\mu_B=m_e/m_p\approx1/1836$ . A proton's magnet is nearly two thousand times feebler than an electron's, so the interaction it produces is correspondingly feeble. The moment sits in the electron's field, and after angular averaging the interaction collapses to one scalar coupling whose eigenstructure is pure angular-momentum algebra — the machinery owned by aq-qm-hyperfine-structure.

\hat H_{hf}=A\,\frac{\hat{\mathbf I}\cdot\hat{\mathbf J}}{\hbar^{2}},\qquad \hat{\mathbf I}\cdot\hat{\mathbf J}=\tfrac12\big[F(F{+}1)-I(I{+}1)-J(J{+}1)\big]\hbar^{2} \qquad (2)

where: $A$ — hyperfine constant, an energy per level packaging $g_I$ and the electronic field at the nucleus; $F$ — total quantum number, integer steps from $|I-J|$ to $I+J$ . The pattern of splitting follows from $\hat{\mathbf I}\cdot\hat{\mathbf J}$ alone; only the scale $A$ waits for the contact calculation. Subtracting the shifts of neighbouring $F$ levels gives the signature spacing — the Landé interval rule, derived in aq-qm-hyperfine-structure — whose growing gaps read off the nuclear spin.

E_F-E_{F-1}=A\,F \qquad (3)

where: the gap between successive $F$ levels grows linearly with the upper $F$ . Measure the ratios of a multiplet's spacings and count $F$ values; with electronic $J$ known, the count fixes $I$ — a classic result of beam spectroscopy that needs no value of $A$ and no wavefunction. For $s$ -states the whole coupling is the Fermi contact term, and its evaluation is what this article owns. The compact form makes every factor visible.

\Delta E(1s)=\frac{4}{3}\,g_p\,\frac{m_e}{m_p}\,\alpha^{4}\,m_ec^{2}=5.87\ \mu\mathrm{eV},\qquad A_s=\frac{2}{3}\mu_0\,g_e\mu_B\,g_p\mu_N\,|\psi(0)|^{2} \qquad (4)

where: $|\psi_{1s}(0)|^{2}=1/(\pi a_0^{3})$ is the ground-state density at the origin; $g_e\approx2.0023$ ; $g_p=5.586$ . Note the compact scaling is $\alpha^{4}m_ec^{2}$ , which equals $(8/3)g_p(m_e/m_p)\alpha^{2}\mathrm{Ry}$ — not the $(4/3)g_p(m_e/m_p)\alpha^{2}\mathrm{Ry}$ that a common misprint gives, a factor of two low. Evaluated in full, the point-proton contact term is $\approx1422$ MHz. The result of that evaluation is one of physics' famous numbers, expressed on every ruler a radio astronomer or a metrologist could want.

\nu_{\mathrm{HFS}}=1420.405751768\ \mathrm{MHz},\quad \lambda=21.106\ \mathrm{cm},\quad \Delta E=5.874\ \mu\mathrm{eV},\quad T=68.2\ \mathrm{mK} \qquad (5)

where: the nine-figure frequency is the measured standard; $\lambda=c/\nu$ is why the line is called " $21$ cm"; $68.2$ mK is $\Delta E/k_B$ , far below the $2.7$ K cosmic microwave background, so both hyperfine levels stay populated everywhere in the cold universe. The same physics one nucleus over defines the second: the caesium ground-state hyperfine transition, chosen for a frequency the 1960s could count and a field-independence the Breit–Rabi formula guarantees.

\nu_{\mathrm{Cs}}=9\,192\,631\,770\ \mathrm{Hz}\ \text{(exact, SI second)},\qquad \lambda=3.26\ \mathrm{cm} \qquad (6)

where: the $^{133}$ Cs $6^{2}S_{1/2}$ level ( $I=7/2$ ) splits into $F=3$ and $F=4$ ; this transition was defined to have exactly this frequency in 1967, so the second is a hyperfine measurement.

Derivation AEvaluating the 21-cm splitting from the Fermi contact term

The hydrogen ground state has $J=S=\tfrac12$ (no orbital field) and proton spin $I=\tfrac12$ , so it couples to $F=1$ or $F=0$ , and the splitting comes entirely from the Fermi contact interaction whose delta-function form is owned by aq-qm-hyperfine-structure. What this card owns is the evaluation: insert the ground-state density at the origin, collect the fundamental constants, and land on the measured $1420$ MHz — then account honestly for the residual that separates a point-proton estimate from the nine-figure standard.

Step 1. Enumerate and evaluate the spins. With $I=J=\tfrac12$ , the coupled states are $F=1$ (triplet, parallel) and $F=0$ (singlet, antiparallel); the spin-dot values are $\hat{\mathbf S}_e\cdot\hat{\mathbf S}_p=+\tfrac14\hbar^{2}$ and $-\tfrac34\hbar^{2}$ , differing by exactly one contact unit. Why the gap is one unit: $\tfrac14-(-\tfrac34)=1$ , so the whole splitting equals the contact constant $A$ with no extra bookkeeping.

Step 2. Insert the density at the origin. The contact energy $\propto|\psi(0)|^{2}$ ; for hydrogen $1s$ , $|\psi_{1s}(0)|^{2}=1/(\pi a_0^{3})$ , the single number that carries the spatial physics. Why only $s$ -states: for $\ell>0$ the centrifugal barrier forces $\psi\propto r^{\ell}$ near the origin, so $\psi(0)=0$ and the contact term switches off.

Step 3. Collect the constants. Grouping $\mu_B\mu_N/a_0^{3}$ into fundamentals gives the compact form $\Delta E=\tfrac43 g_p\tfrac{m_e}{m_p}\alpha^{4}m_ec^{2}$ . With $g_p=5.586$ , $m_e/m_p=1/1836$ , $\alpha=1/137.036$ , $m_ec^{2}=0.511$ MeV this is $5.877\times10^{-6}$ eV. Why the power is $\alpha^{4}$ , not $\alpha^{2}$ : two powers come from $\mu_B\mu_N\propto\alpha^{2}$ -scale magnetism and two from $1/a_0^{3}\propto\alpha^{3}$ against $m_ec^{2}$ — the honest count is $\alpha^{4}m_ec^{2}=(8/3)g_p(m_e/m_p)\alpha^{2}\mathrm{Ry}$ , and dropping to $(4/3)\alpha^{2}\mathrm{Ry}$ halves the answer.

Step 4. Convert and compare. $\nu=\Delta E/h=1.42\times10^{9}$ Hz $=1420$ MHz, $\lambda=c/\nu=21.1$ cm; the full point-proton value with $g_e=2.0023$ is $\approx1422$ MHz, a part in $10^{3}$ above the measured $1420.406$ MHz. Why the residual: reduced mass, bound-state QED, and the proton's finite magnetization radius (the Zemach correction) — the last digits are a precision story, not a flaw in the contact picture. *Verified in verify_aq-optics-hyperfine-structure.py.*

Go deeper · Derivation B · 1 minThe 21-cm line: why a forbidden transition still shines

Derivation BThe 21-cm line: why a forbidden transition still shines

The $F=1\to F=0$ transition is magnetic-dipole (M1), the slow cousin of the electric-dipole decays that empty ordinary excited states, and its spontaneous rate is minuscule. Yet it is the brightest line in radio astronomy. This card computes the rate and lifetime and explains the resolution of the paradox: an eleven-million-year lifetime is irrelevant when the emitters number $10^{60}$ per sightline and collisions keep the upper level filled.

Step 1. The rate. The magnetic-dipole spontaneous coefficient is $A_{21}=\mu_0\omega^{3}|\langle\mu\rangle|^{2}/(3\pi\hbar c^{3})$ with $\omega=2\pi\nu_{\mathrm{HFS}}$ and a moment of order $\mu_B$ ; evaluated, $A_{21}\approx2.85\times10^{-15}\ \mathrm{s}^{-1}$ . Why it is so tiny: M1 rates scale as $\omega^{3}$ , and a $1.4$ GHz frequency is roughly $10^{6}$ times below an optical line, cubing to a $10^{18}$ suppression on top of the M1 selection.

Step 2. The lifetime. Inverting, $\tau=1/A_{21}\approx1.1\times10^{7}$ years — about $11$ million years for a single atom to flip. Why a lab never sees it: no terrestrial sample is undisturbed for megayears, so the line is invisible on Earth and only the Galaxy provides the path length.

Step 3. Why it dominates the radio sky. A galactic sightline holds of order $10^{20}$ hydrogen atoms per square centimetre; even with a $10^{-15}\ \mathrm{s}^{-1}$ rate the summed emission is enormous, and collisions plus the $2.7$ K background continually repopulate $F=1$ . Why abundance beats probability: the line strength is a rate times a population, and interstellar hydrogen supplies population without limit. *Verified in verify_aq-optics-hyperfine-structure.py.*

Go deeper · Derivation C · 1 minThe caesium clock and the field-immune transition

Derivation CThe caesium clock and the field-immune transition

The SI second rests on the caesium ground-state hyperfine transition, and the reason a hyperfine line makes a good clock is field-independence, not just a convenient frequency. This card locates the $F=3\leftrightarrow F=4$ interval, shows why the $m_F=0\to m_F=0$ pair is chosen, and connects that immunity to the Breit–Rabi field dependence owned by aq-qm-hyperfine-structure.

Step 1. The transition. Caesium-133 has $I=7/2$ ; its $6^{2}S_{1/2}$ ground level splits into $F=3$ and $F=4$ , and the interval was fixed by definition at $9\,192\,631\,770$ Hz in 1967 — a $3.26$ cm microwave. Why $9.2$ GHz: it was, in the 1960s, a frequency that could be counted directly against the era's electronics, unlike an optical line.

Step 2. The field-immune pair. The Breit–Rabi formula gives each $m_F$ level's energy as a function of $B$ ; the $m_F=0$ states of $F=3$ and $F=4$ have zero first derivative in $B$ at $B=0$ , so their separation is flat there. Why it matters: a stray magnetic field shifts the clock transition only quadratically, so the reference is exceptionally reproducible.

Step 3. The fountain. Cold caesium atoms tossed upward through a microwave cavity interrogate the $m_F=0\to0$ line twice (Ramsey separated oscillations), and the quadratic-only field dependence lets the residual Zeeman shift be measured and removed to $10^{-16}$ . Why it closes: the same immunity that makes the transition dull in a magnetometer makes it superb in a clock. *Verified in verify_aq-optics-hyperfine-structure.py.*

Limiting cases

Case	Result	Meaning
$I=0$ (¹²C, ¹⁶O)	no hyperfine structure	no nuclear magnet to couple; $F=J$ trivially
$s$ -state ( $\ell=0$ )	Fermi contact, $\Delta E\propto\lvert\psi(0)\rvert^{2}\propto Z^{3}/n^{3}$	only $s$ -states reach the nucleus; largest hyperfine, source of the 21 cm line
$\ell>0$	dipole–dipole plus orbital term, no contact	$\psi(0)=0$ ; hyperfine from the distant field, smaller
weak field $\mu_B B\ll A$	$F,m_F$ good; $2F+1$ fan, slope $g_F\mu_B m_F$	hyperfine wins; the hyperfine Zeeman effect
strong field $\mu_B B\gg A$	$m_I,m_J$ good; $E\approx g_J\mu_B B\,m_J+A\,m_I m_J$	Back–Goudsmit decoupling — $\mathbf I$ and $\mathbf J$ precess separately
$I\ge1$	add electric-quadrupole term	non-spherical nucleus; the interval rule acquires a correction

Common trapHyperfine structure is just fine structure done again

No — the mechanism is different and so is the reason it is small. Fine structure is relativistic and spin-orbit, an $O(\alpha^{2})$ correction to the Bohr energy; hyperfine structure is the nuclear magnet, suppressed by the extra factor $\mu_N/\mu_B=m_e/m_p\approx1/1836$ — a mass ratio, not another power of $\alpha$ , which lands it a further $\sim g_p\,m_e/m_p\approx1/300$ below fine structure. The good quantum number is genuinely new: $\mathbf F=\mathbf I+\mathbf J$ includes nuclear spin, which never appeared in the fine-structure problem. And the $21$ -cm transition is magnetic-dipole, not electric-dipole — that is why its lifetime is $11$ million years rather than nanoseconds, and why it took an accident of cosmic abundance to make it the most-mapped line in the sky.

Worked examples

Example 1 — 21-cm Doppler galactography (the observatory number). A hydrogen cloud recedes along the line of sight at $v_{\mathrm{LOS}}=100\ \mathrm{km/s}$ . Find the Doppler shift of its $21$ -cm line and say why it maps the Galaxy.

Step 1. The shift. First-order Doppler gives $\Delta\nu=\nu_{\mathrm{HFS}}\,v/c = 1420.406\times10^{6}\times(10^{5}/2.998\times10^{8})$ .

Step 2. Evaluate. $\Delta\nu = 1420.406\times10^{6}\times3.336\times10^{-4} = \mathbf{473.8\ kHz}$ — a shift radio receivers resolve to hertz.

Step 3. The payoff. Because $\Delta\nu/\nu\approx3\times10^{-4}$ maps linearly to velocity, every cloud's $21$ -cm frequency reports its line-of-sight speed to better than $1\ \mathrm{km/s}$ ; assembling these across the sky yields the Milky Way's rotation curve — the flat curve that first weighed the dark halo. Answer: $\Delta\nu=473.8$ kHz at $100$ km/s, so hertz-level radio spectroscopy reads galactic velocities to km/s.

Example 2 — The energy pyramid and the spin temperature. Rank hydrogen's small splittings, then ask how the $2.7$ K cosmos populates the $21$ -cm levels.

Step 1. The ladder. Fine structure $2P$ : $45.3\ \mu$ eV; hyperfine $1s$ : $5.874\ \mu$ eV; Lamb $2s$ : $4.375\ \mu$ eV (the subject of The Lamb Shift). So hyperfine is about $0.13$ of fine structure and $1.34$ times the Lamb shift — three rungs within a decade.

Step 2. The temperature. Converting the hyperfine gap, $\Delta E/k_B = h\nu_{\mathrm{HFS}}/k_B = \mathbf{68.2\ mK}$ , far below the $2.725$ K microwave background.

Step 3. The populations. With statistical weights $3:1$ and $k_BT_{\mathrm{CMB}}\gg\Delta E$ , the Boltzmann ratio is $n(F{=}1)/n(F{=}0)=3\,e^{-\Delta E/k_BT}\approx 3\times0.975 = 2.93$ — essentially the full degeneracy ratio. Answer: the gap is $68.2$ mK, so the CMB keeps both levels populated at nearly $3:1$ , and the "spin temperature" of the gas is what $21$ -cm surveys actually measure.

Physics of the portrait

Hyperfine structure is a concept assembled by a relay of people across half a century — from a proposal to explain crowded spectral lines to the horn antenna that heard hydrogen sing.

Origins · 1 minhyperfine structure and how this was discovered

hyperfine structure

reading the nucleus in the spectrum · Pauli 1924 → Fermi 1930 → Rabi 1938 → van de Hulst 1944 → Ewen and Purcell 1951

Wolfgang Pauli proposed in 1924 that the faint, crowded splitting of certain spectral lines meant the nucleus carried spin and a magnetic moment — an audacious guess that made the nucleus a quantum object with angular momentum. Enrico Fermi, in Rome in 1930, turned the guess into a formula: he showed that the naively divergent dipole field inside the nuclear volume yields a finite contact term proportional to $|\psi(0)|^{2}$ , nonzero only for $s$ -states — the reason the hydrogen ground state splits at all. Isidor Rabi built the measuring instrument: his 1938 molecular-beam magnetic-resonance method read nuclear moments straight off a deflected beam, founding the resonance techniques that became NMR, MRI, and the atomic clock (Nobel 1944). Hendrik van de Hulst, a Dutch student under wartime occupation in 1944, predicted that the $1s$ hyperfine transition would radiate a detectable line at $21$ cm from interstellar hydrogen — a line no one had sought because its lifetime seemed impossibly long. Harold Ewen and Edward Purcell found it in 1951 with a plywood horn antenna bolted outside a Harvard laboratory window; Dutch and Australian groups confirmed it within weeks, and radio astronomy of neutral hydrogen was born. Each name owns a rung: the spin, the contact term, the beam, the prediction, the detection.

History

1924

— Pauli proposes nuclear spin to explain the hyperfine splitting of spectral lines

1930

— Fermi derives the s-state contact interaction proportional to |ψ(0)|²

1931

— Breit and Rabi give the closed-form field dependence of the J = 1/2 ground state

1938

— Rabi's molecular-beam magnetic resonance reads nuclear moments off the beam

1944

— van de Hulst predicts a detectable 21 cm line from interstellar hydrogen

1951

— Ewen and Purcell detect the 21 cm line from a horn antenna at Harvard

1960

— Goldenberg, Kleppner, and Ramsey run the first hydrogen maser

1967

— the caesium hyperfine transition, 9 192 631 770 Hz, is adopted as the SI second

Going further

Texts. Foot, Atomic Physics, Ch. 6 (hyperfine structure and the interval rule) · Demtröder, Laser Spectroscopy, Vol. 1, §2.4 · Bransden and Joachain, Physics of Atoms and Molecules, Ch. 5 · Woodgate, Elementary Atomic Structure, Ch. 9.

Provenance. The coupling $A\,\hat{\mathbf I}\cdot\hat{\mathbf J}$ , the $\hat{\mathbf I}\cdot\hat{\mathbf J}$ eigenvalue, the interval rule, the Fermi contact term, and the Breit–Rabi formula are recalled by name from aq-qm-hyperfine-structure and are not re-derived here; this article owns the evaluation, the $21$ -cm astronomy, the caesium clock, and the beam spectroscopy. Base provenance preserved verbatim: measured line $1420.405751768$ MHz and caesium clock $9\,192\,631\,770$ Hz (defined) are standard values; $g_p\approx5.5857$ , $m_p/m_e\approx1836.15$ (CODATA 2018). A verifier should confirm the $\hat{\mathbf I}\cdot\hat{\mathbf J}$ eigenvalue and interval rule symbolically, $|\psi_{1s}(0)|^{2}=1/(\pi a_0^{3})$ , the contact evaluation to $\approx1422$ MHz, the $A_{21}$ rate and $11$ Myr lifetime, and the Doppler and spin-temperature numbers — all executed in verify_aq-optics-hyperfine-structure.py. Edition note (base repair): the dry base wrote the compact contact splitting as $\tfrac43 g_p(m_e/m_p)\alpha^{2}\mathrm{Ry}$ , which is a factor of two too small (it yields $\approx710$ MHz); the correct form is $\tfrac43 g_p(m_e/m_p)\alpha^{4}m_ec^{2}=\tfrac83 g_p(m_e/m_p)\alpha^{2}\mathrm{Ry}=5.87\ \mu$ eV $=1420$ MHz — corrected here and flagged in the derivation and cheat sheet.

Prerequisites. Fine Structure and Spin-Orbit Coupling · Electron Spin and the Spin Magnetic Moment · Hydrogenic Wavefunctions and Radial Distributions · Quantum Numbers and Orbital Angular Momentum. Leads to. The Zeeman Effect · The Paschen-Back and Anomalous Zeeman Regimes · The Lamb Shift · The 21-cm Hydrogen Line.

Why does the nuclear spin matter at all — isn't the nucleus just a point charge?

For gross and fine structure it effectively is. But the proton is a spin- $\tfrac12$ particle with a magnetic moment, and a magnet in the electron's field has an orientation energy — resolved by whether the two moments are parallel or antiparallel, that energy is hyperfine structure. It is tiny because the nuclear moment is $\sim1/1836$ of the electron's, but it is not zero, and for $s$ -states it produces the $21$ -cm line that maps the Galaxy.

Where does the interval rule E_F − E_{F−1} = A·F come from?

Purely from $\hat{\mathbf I}\cdot\hat{\mathbf J}=\tfrac12[F(F+1)-I(I+1)-J(J+1)]\hbar^{2}$ , the eigenvalue owned by aq-qm-hyperfine-structure. The energy of level $F$ is $\tfrac A2$ times that bracket; subtract adjacent $F$ and only the $F(F+1)$ term survives, leaving $A\,F$ . No wavefunction and no value of $A$ enter, which is exactly why measuring the spacing ratios in a multiplet determines the nuclear spin $I$ directly.

If only s-states feel the contact term, do p-states have no hyperfine structure?

They do — just not the contact kind. For $\ell>0$ the density vanishes at the origin, so the contact term is off, but the electron's orbital current and its distant dipole field still couple to the nuclear moment. That dipolar-plus-orbital hyperfine interaction is generally smaller than a comparable contact term and produces the interval-rule multiplets of levels like sodium's $3^{2}P_{3/2}$ , whose four sublevels space as $1:2:3$ and pin $^{23}$ Na's spin at $I=3/2$ .

Why does caesium define the second rather than an optical line?

Two reasons. First, the $m_F=0$ clock transition is field-independent to first order (the flat Breit–Rabi curve), so a stray magnetic field barely shifts it — exceptional reproducibility. Second, $9.2$ GHz was, in 1967, a frequency the era's electronics could count directly. Optical clocks are now far more precise, and a redefinition of the second in terms of an optical transition is under active discussion — but the principle stays the same hyperfine physics.

Is the 5.87 µeV I computed the "real" 21 cm energy?

It is the leading contact estimate, right to a part in $10^{3}$ . The measured line is $1420.405751768$ MHz — nine significant figures — and the gap between the two is reduced-mass, bound-state QED, and finite-proton-size (Zemach) corrections plus the fact that the proton is not a point dipole. The simple formula gets the physics — the $g_p$ , the $m_e/m_p$ , the $\alpha^{4}$ , the $|\psi(0)|^{2}$ — exactly right; the last digits are a precision-QED story that The 21-cm Hydrogen Line carries.

How can an 11-million-year lifetime give a bright astronomical line?

Brightness is a rate times a population, and interstellar hydrogen supplies population without limit — of order $10^{20}$ atoms per square centimetre along a galactic sightline. Even a spontaneous rate of $10^{-15}\ \mathrm{s}^{-1}$ summed over that many atoms yields copious emission, and collisions plus the microwave background keep the upper level filled. The long lifetime only means the line is invisible in any earthbound sample; give it a galaxy's worth of path length and it becomes the most-mapped line in radio astronomy.