Mössbauer Effect

The problem: recoil destroys nuclear resonance

Resonant absorption is everywhere in atomic physics. A sodium atom emits a 589 nm photon and another sodium atom soaks it up — the emission and absorption lines sit at the same energy, so the cross-section is enormous. The natural expectation is that the same trick works for nuclei: a nucleus drops from an excited state, emits a gamma ray, and an identical nucleus next door absorbs it back into the same excited state.

For nuclei it almost never works for free atoms, and the culprit is recoil. Conservation of momentum forces the emitting nucleus to kick backward when the photon leaves, exactly like a rifle recoiling from a bullet. A nucleus of mass M emitting a photon of energy E_γ recoils with momentum p = E_γ/c and carries away kinetic energy:

E_R = p² / (2M) = E_γ² / (2 M c²)

That recoil energy comes out of the photon, so the emitted gamma is below the true transition energy by E_R. Worse, the absorbing nucleus must also recoil when it catches the photon, so it needs a photon above the transition energy by E_R. Emission and absorption lines are therefore separated by 2E_R, and on top of that thermal motion smears each line by the Doppler width.

Plug in numbers for the workhorse transition, the 14.4 keV line of iron-57. With M ≈ 57 atomic mass units:

E_R = (14400 eV)² / (2 × 57 × 931.5×10⁶ eV) ≈ 1.95 × 10⁻³ eV

Now compare that to the natural linewidth. The 14.4 keV state of ⁵⁷Fe lives for a mean lifetime τ ≈ 141 ns (half-life ≈ 98 ns), and the Heisenberg energy–time relation Γ = ħ/τ gives a width of only about 4.7 × 10⁻⁹ eV. The recoil shift is roughly 10⁶ linewidths. The emission and absorption lines miss each other by a million widths — there is essentially zero overlap, and nuclear gamma resonance is dead in the water.

Quantity (⁵⁷Fe 14.4 keV)	Value	In units of Γ
Natural linewidth Γ	≈ 4.7 × 10⁻⁹ eV	1
Free-atom recoil E_R	≈ 1.95 × 10⁻³ eV	≈ 4 × 10⁵
Doppler width at 300 K	≈ 1 × 10⁻² eV	≈ 2 × 10⁶
Relative resolution Γ/E_γ	≈ 3 × 10⁻¹³	—

Mössbauer's insight: hand the recoil to the whole crystal

In 1957, working on his doctoral thesis, Rudolf Mössbauer was studying gamma resonance in iridium-191 and expected the recoil and thermal Doppler broadening to weaken as he cooled the source. To everyone's surprise the resonance got stronger. The explanation he gave is the heart of the effect.

When the emitting nucleus is bound in a solid, the recoil momentum does not necessarily go into kicking that one atom loose. Quantum mechanically, the lattice can only accept energy in discrete vibrational quanta — phonons. There is a finite probability that a gamma is emitted creating zero phonons. In that case the recoil momentum is taken up by the rigid lattice as a whole, whose mass is ~10²⁰ times that of one nucleus. The recoil energy E_R = E_γ²/(2 M_crystal c²) becomes utterly negligible, and the photon leaves with the full transition energy. The line is recoil-free, has no Doppler thermal broadening, and is as sharp as the natural linewidth allows.

The fraction of emissions that happen this way is the recoil-free fraction f, also called the Lamb–Mössbauer factor:

f = exp(−k²⟨x²⟩) = exp(−E_γ²⟨x²⟩ / (ħ²c²))

Here k = E_γ/(ħc) is the gamma wavenumber and ⟨x²⟩ is the mean-square displacement of the nucleus along the photon direction due to lattice vibrations. The physics packed into this exponent is intuitive once unpacked:

• Low gamma energy helps. f falls off as exp(−E_γ²). Doubling the gamma energy quarters the exponent's friendliness, which is why the effect favours soft transitions like 14.4 keV (⁵⁷Fe) and 23.9 keV (¹¹⁹Sn) and is hopeless for MeV gammas.
• Stiff lattices help. A high Debye temperature θ_D means small ⟨x²⟩, so f is large. Loosely bound atoms or liquids have huge ⟨x²⟩ and f ≈ 0 — there is no Mössbauer effect in a gas or liquid.
• Cold helps. Cooling reduces thermal vibration, shrinking ⟨x²⟩ and raising f, which is exactly the counter-intuitive temperature trend Mössbauer first observed.

In the Debye model ⟨x²⟩ has a zero-point part that survives even at absolute zero, so f never quite reaches 1. For metallic iron at room temperature f ≈ 0.7–0.8; for tin-119 it is lower and benefits from cooling; for heavy nuclei like dysprosium the recoil-free fraction is so small at 300 K that the experiment must run near liquid-helium temperature.

Just how sharp is the line?

The ratio that matters for any spectroscopy is the relative linewidth Γ/E_γ. For ⁵⁷Fe:

Γ / E_γ ≈ (4.7 × 10⁻⁹ eV) / (1.44 × 10⁴ eV) ≈ 3 × 10⁻¹³

That is the highest energy resolution of any everyday spectroscopic technique — you can resolve an energy change of a few parts in 10¹³. To put it in motion units, the Doppler shift ΔE = (v/c)·E_γ equal to one linewidth corresponds to a source velocity of about 0.1 mm/s. The full hyperfine structure of iron compounds is spread over velocities of only ±10 mm/s — slower than a snail. This is why a Mössbauer drive is literally a loudspeaker coil nudging the radioactive source back and forth at millimetres per second; that gentle motion is enough to Doppler-scan the photon energy across the entire resonance.

Mössbauer spectroscopy and hyperfine interactions

A recoil-free, ultra-narrow gamma line is a ruler with 10¹³ tick marks. Any interaction that nudges the nuclear energy levels by even a fraction of an eV shows up. In a typical transmission setup a ⁵⁷Co source (which decays to the excited 14.4 keV state of ⁵⁷Fe) sits on a velocity drive; gamma rays pass through an iron-containing sample and into a detector. As the source velocity sweeps, the transmitted intensity dips wherever the Doppler-shifted photon hits a resonance in the sample. The pattern of dips encodes three hyperfine interactions:

Hyperfine interaction	Physical origin	Spectral signature
Isomer (chemical) shift δ	Electric monopole: s-electron density at the nucleus differs between source and sample	Whole spectrum shifts by ~0.1–2 mm/s; tracks oxidation state and bonding
Quadrupole splitting ΔE_Q	Nuclear quadrupole moment couples to the electric field gradient of an asymmetric charge environment	Single line splits into a doublet
Magnetic hyperfine splitting	Nuclear magnetic moment in the internal magnetic field at the nucleus (nuclear Zeeman effect)	Sextet of lines; spacing gives the hyperfine field B (≈ 33 T in α-iron)

From these three observables alone, a Mössbauer spectrum reveals whether iron is Fe²⁺ or Fe³⁺, whether it is high-spin or low-spin, whether the material is magnetically ordered, and at what temperature it loses that order. Because the technique reports on the nucleus itself, it is element-specific and even isotope-specific — it sees only the ⁵⁷Fe in the sample and ignores everything else.

Weighing photons: the Pound–Rebka experiment

The most famous application is not chemistry at all but gravity. General relativity predicts that a photon climbing out of a gravitational potential loses energy — the gravitational redshift. Over a height h on Earth the fractional shift is:

Δf / f = g·h / c²

For the 22.5-metre tower in the Jefferson Laboratory at Harvard, g·h/c² ≈ 2.5 × 10⁻¹⁵. No ordinary clock or spectrometer could see a shift that small in 1959 — but the ⁵⁷Fe Mössbauer line, with Γ/E_γ ≈ 3 × 10⁻¹³, was within reach. Robert Pound and Glen Rebka sent 14.4 keV gammas up and down the tower and used a slow drive velocity to cancel the gravitational shift, reading off the redshift from the velocity needed. Their result matched Einstein's prediction to about 10 percent (refined to ~1 percent by Pound and Snider in 1965) — the first terrestrial confirmation of gravitational redshift, and a demonstration that the Mössbauer line is sharp enough to literally weigh a photon's gravitational energy.

Worked numbers

Scenario	Result
⁵⁷Fe 14.4 keV free-atom recoil	E_R ≈ 1.95 × 10⁻³ eV (≈ 4 × 10⁵ linewidths)
Velocity for one-linewidth Doppler shift	v = c·Γ/E_γ ≈ 0.10 mm/s
Velocity to scan full α-iron sextet	≈ ±11 mm/s
Hyperfine field in metallic α-iron	≈ 33 T (gives ~10.6 mm/s outer-line splitting)
Pound–Rebka gravitational shift (22.5 m)	Δf/f ≈ 2.5 × 10⁻¹⁵ → v ≈ 7.5 × 10⁻⁷ m/s
Recoil-free fraction, α-Fe at 300 K	f ≈ 0.7–0.8

JavaScript — recoil and Mössbauer arithmetic

// Constants
const c = 2.998e8;            // m/s
const amu = 931.494e6;        // eV/c^2 per atomic mass unit
const hbar_eVs = 6.582e-16;   // eV·s

// Free-atom recoil energy E_R = E_gamma^2 / (2 M c^2)
function recoilEnergy(Egamma_eV, massAmu) {
  return (Egamma_eV * Egamma_eV) / (2 * massAmu * amu);
}

// Natural linewidth from mean lifetime: Gamma = hbar / tau
function linewidth(tau_s) {
  return hbar_eVs / tau_s;
}

// Iron-57: 14.4 keV, mass 57 amu, mean lifetime ~141 ns
const Eg = 14400;                       // eV
const ER = recoilEnergy(Eg, 57);        // ~1.95e-3 eV
const Gamma = linewidth(141e-9);        // ~4.7e-9 eV
console.log(`Recoil energy:   ${ER.toExponential(2)} eV`);
console.log(`Linewidth Gamma: ${Gamma.toExponential(2)} eV`);
console.log(`Recoil = ${(ER / Gamma).toExponential(1)} linewidths`); // ~4e5

// Relative resolution Gamma / E_gamma
console.log(`Resolution: ${(Gamma / Eg).toExponential(1)}`);  // ~3e-13

// Doppler velocity that shifts the photon by one linewidth
// dE = (v/c) * E_gamma  ->  v = c * dE / E_gamma
function velocityForShift(dE_eV, Egamma_eV) {
  return c * dE_eV / Egamma_eV;   // m/s
}
console.log(`v for 1 linewidth: ${(velocityForShift(Gamma, Eg) * 1000).toFixed(3)} mm/s`);

// Recoil-free fraction (Lamb-Mossbauer): f = exp(-k^2 )
// k = E_gamma / (hbar c). Pass mean-square displacement in m^2.
function recoilFreeFraction(Egamma_eV, x2_m2) {
  const k = Egamma_eV / (hbar_eVs * c);   // 1/m
  return Math.exp(-k * k * x2_m2);
}
// Typical  for alpha-Fe near room T ~ 5e-23 m^2
console.log(`f (alpha-Fe, 300K): ${recoilFreeFraction(Eg, 5e-23).toFixed(2)}`);

// Pound-Rebka gravitational redshift over height h
function gravRedshift(h_m, g = 9.81) {
  return g * h_m / (c * c);   // fractional Df/f
}
const dff = gravRedshift(22.5);
console.log(`Df/f over 22.5 m: ${dff.toExponential(2)}`);              // ~2.5e-15
console.log(`Equivalent v: ${(dff * c * 1e9).toFixed(0)} nm/s`);       // ~750 nm/s

Where the Mössbauer effect shows up

• Mineralogy and geology. Iron oxidation states and magnetic phases in rocks; NASA's Spirit and Opportunity rovers carried MIMOS II Mössbauer spectrometers to identify hematite, magnetite, and olivine on Mars.
• Biochemistry. Iron centres in haemoglobin, ferredoxins, and nitrogenase; ⁵⁷Fe-enriched proteins reveal spin and oxidation state of active sites.
• Magnetism. Measuring internal hyperfine fields, Curie and Néel temperatures, and spin structures in alloys, oxides, and nanoparticles.
• Corrosion and materials science. Identifying rust phases, steel microstructures, and amorphous-to-crystalline transitions.
• Fundamental physics. Gravitational redshift (Pound–Rebka), tests of special relativity (transverse Doppler / time dilation via rotor experiments), and searches for tiny symmetry violations.
• Nuclear physics. Measuring nuclear quadrupole and magnetic moments, and lifetimes of low-lying excited states.

Common misconceptions

• "The recoil energy goes away." It doesn't vanish — in a recoil-free event it goes into the entire crystal's translational motion, where E_R = E_γ²/(2M_crystal c²) is negligible because the mass is enormous. Energy and momentum are both conserved.
• "It works for any gamma transition." No. The recoil-free fraction dies off as exp(−E_γ²), so only low-energy transitions (tens of keV) in stiff solids give a usable effect. MeV gammas have f ≈ 0.
• "It happens in liquids and gases too." No. Recoil-free emission requires the nucleus to be rigidly bound. In a liquid or gas ⟨x²⟩ is huge and f ≈ 0; you must be in a solid lattice.
• "Every emission is recoil-free." Only a fraction f are. The remaining (1−f) emissions create phonons and produce a broad, useless recoil-shifted background; the sharp line sits on top of it.
• "The isomer shift is a Doppler effect." The isomer shift is a real change in nuclear energy levels from differing s-electron density between source and absorber; it is simply measured using the Doppler drive, not caused by it.
• "Higher source velocity means better resolution." The resolution is set by the natural linewidth Γ. The drive velocity merely tunes the photon energy; a few mm/s already covers the full hyperfine pattern.